Surface Chemistry

SPRINGER BRIEFS IN MATERIALS
E. McCafferty
Surface
Chemistry
of Aqueous
Corrosion
Processes
123
SpringerBriefs in Materials
More information about this series at http://www.springer.com/series/10111
E. McCafferty
Surface Chemistry
of Aqueous Corrosion
Processes
123
E. McCafferty
Alexandria, VA
USA
ISSN 2192-1091 ISSN 2192-1105 (electronic)

SpringerBriefs in Materials
ISBN 978-3-319-15647-7 ISBN 978-3-319-15648-4 (eBook)
DOI 10.1007/978-3-319-15648-4
Library of Congress Control Number: 2015932239
Springer Cham Heidelberg New York Dordrecht London

© The Author(s) 2015
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained
herein or for any errors or omissions that may have been made.
Printed on acid-free paper
Springer International Publishing AG Switzerland is part of Springer Science+Business Media

(www.springer.com)
Preface
This communication utilizes a surface chemistry/physical chemistry approach

toward the study of aqueous corrosion processes. Topics included are acid–base
properties of surface oxide films, capillarity and corrosion, corrosion inhibition,
metal/polymer adhesion, formation of water films on the iron oxide surface, pas-
sivity of Fe–Cr and Fe–Cr–Ni alloys, the relationship between the isoelectric point
and the potential of zero charge for passive metals, and the uptake of chloride ions
and the pitting of aluminum.
Alexandria, VA, USA, 2015 E. McCafferty
v
Contents
1 Acid-Base Properties of Surface Oxide Films . . . . . . . . . . . . . . . . . 1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Surface Hydroxyl Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Nature of Acidic and Basic Surface Sites . . . . . . . . . . . . . . . . . 3
Determination of Isoelectric Points of Metal Oxides
and Oxide Films . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Streaming Potential Measurement . . . . . . . . . . . . . . . . . . . . . . . 5
Contact Angle Measurements. . . . . . . . . . . . . . . . . . . . . . . . . . 5
Surface Reaction Method of Simmons and Beard . . . . . . . . . . . . 8
Ionic Interaction Model of Parks . . . . . . . . . . . . . . . . . . . . . . . 10
XPS Correlation Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Measurement of Near-Surface Forces . . . . . . . . . . . . . . . . . . . . 15
Numerical Values of Isoelectric Points . . . . . . . . . . . . . . . . . . . 15
Surface Charge of Oxide Films . . . . . . . . . . . . . . . . . . . . . . . . 16
Pitting of Aluminum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Adsorption of Cl− on Aluminum . . . . . . . . . . . . . . . . . . . . . . . 17
Electrode Kinetics of Pit Initiation on Aluminum . . . . . . . . . . . . 18
Pitting Potential of Aluminum as a Function of PH . . . . . . . . . . 20
Pitting Potential of Aluminum as a Function of Cl−
Concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
Acid-Base Interactions in Metal/Polymer Adhesion . . . . . . . . . . . . . . 23
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Theories of Polymer Adhesion . . . . . . . . . . . . . . . . . . . . . . . . . 23
Acid-Base Nature of Organic Polymers . . . . . . . . . . . . . . . . . . . 24
Characterization of the Acid-Base Properties of Polymers . . . . . . 25
Metal/Polymer Adhesion Tests. . . . . . . . . . . . . . . . . . . . . . . . . 28
Wet Adhesion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Acid-Base Effects in Corrosion Inhibition . . . . . . . . . . . . . . . . . . . . . 31
General Effect of Oxide Films . . . . . . . . . . . . . . . . . . . . . . . . . 31
Mechanisms of Interaction with Oxide Films . . . . . . . . . . . . . . . 33
Interaction of Inhibitors with Hydroxylated Oxide Films . . . . . . . 33
vii
viii Contents
Chelating Compounds as Corrosion Inhibitors . . . . . . . . . . . . .. 33

Hard and Soft Acids and Bases . . . . . . . . . . . . . . . . . . . . . . .. 35
Interaction of Cations with Oxide Films . . . . . . . . . . . . . . . . .. 38
Relationship Between the Isoelectric Point (pHpzc) and the Potential
of Zero Charge (Epzc) for Passive Metals . . . . . . . . . . . . . . . . . . . . . 39
General Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Measurement of Potentials of Zero Charge for Passive Metals . . . 41
Potential Drop in the Metal/Oxide/Electrolyte System . . . . . . . . . 41
Space Charge Side of the Oxide/Solution Interface . . . . . . . . . . . 42
Solution Side of the Oxide/Solution Interface. . . . . . . . . . . . . . . 44
Numerical Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2 Capillarity and Corrosion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Capillarity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Application to Organic Coatings . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
Young and Laplace Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
Force Between Glass Plates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Application to Poultice Corrosion . . . . . . . . . . . . . . . . . . . . . . . . . . 60
Application to White Rusting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3 A Recent Model of Passivity for Fe-Cr and Fe-Cr-Ni Alloys. . . . . . 63

Fe-Cr Alloys . . . . . . . . . . . . . . . ..................... . . . . . 63
Fe-Cr-Ni Alloys. . . . . . . . . . . . . ..................... . . . . . 66
References . . . . . . . . . . . . . . . . ..................... . . . . . 69
4 Uptake of Chloride Ions and the Pitting of Aluminum . . . . . . . . . . 71

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Sequence of Steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
XPS Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Formation of Blisters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Localized Corrosion Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
5 Formation of Water Films on the Iron Oxide Surface . . . . . . . . . . 77

Introduction . . . . . . . . . . . . . . . . ............... . . . . . . . . . . 77
Experimental . . . . . . . . . . . . . . . . ............... . . . . . . . . . . 77
Results and Discussion . . . . . . . . . ............... . . . . . . . . . . 78
Formation of Liquid-Like Layers . . ............... . . . . . . . . . . 81
References . . . . . . . . . . . . . . . . . ............... . . . . . . . . . . 83
Contents ix
6 Corrosion Inhibition by Fluorinated Aliphatic Compounds . . . . . . 85

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
Fluorinated Carboxylic Acids and Amines . . . . . . . . . . . . . . . . . . . . 85
Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Suggestions for Further Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Chapter 1
Acid-Base Properties of Surface
Oxide Films
Introduction
The properties of metal oxide films are important in a number of corrosion phe-
nomena, including passivity and its breakdown, corrosion inhibition by organic
molecules, and the adhesion of organic coatings. In addition, the properties of metal
oxide films are important in related surface phenomena, such as friction and wear,
lubrication, and electrocatalysis. This chapter will consider the acid-base nature of
oxide films and the role of acid-base properties in pitting, the adhesion of organic
coatings, and corrosion inhibition. Examples and references given here are intended
to be illustrative rather than being exhaustive.
Surface Hydroxyl Groups
It is well known that oxide surfaces or oxide-covered metals exposed to either the
ambient environment or immersed in aqueous solutions terminate in an outermost
layer of hydroxyl groups due to their interaction with water molecules [1–3]. The
reaction of a surface oxide with a water molecule is depicted in Fig. 1.1. Using
quantitative X-ray photoelectron spectroscopy, the concentration of surface
hydroxyls in the air-formed oxide films on various metals has been found to be
6–20 OH groups/nm2 [4]. The depth of the hydroxylated region extends 5–8 Å into
the oxide. Although the surface hydroxyl layer is quite thin, it has nonetheless a
large influence on the properties of the oxide film, especially on its surface charge.
In aqueous solutions, the surface hydroxyl groups may remain undissociated. In
such an instance, the oxide surface is said to be at its isoelectric point (pHpzc) and
will have a net surface charge of zero. This will occur only if the pH of the aqueous
solution has the same value as the isoelectric point of the oxide. It is more likely,
© The Author(s) 2015 1

E. McCafferty, Surface Chemistry of Aqueous Corrosion Processes,
SpringerBriefs in Materials, DOI 10.1007/978-3-319-15648-4_1
2 1 Acid-Base Properties of Surface Oxide Films
Fig. 1.1 Interaction of a water molecule with a metal oxide surface to form an outermost layer of
surface hydroxyl groups
however, that the oxide film will be charged. If the pH is less than the isoelectric
point, the surface will acquire a positive charge:
MOHsurf þ Hþ ðaqÞ MOH2 þ surf ð1:1Þ
where M denotes a surface site occupied by a metal cation. If the pH is greater than
the isoelectric point, the surface will acquire a negative charge:
MOHsurf þ OH MO surf þ H2 O
or:
MOHsurf MO surf þ Hþ ðaqÞ ð1:2Þ
In the above reactions the surface species MOHþ 2 is a Bronsted acid because it
is a proton donor [in the reverse direction of Eq. (1.1)]; and −MO− is a Bronsted
base in Eq. (1.2) (reverse direction) because it is a proton acceptor. The surface
species −MOH is amphoteric, being a Bronsted base in Eq. (1.1) (forward direc-
tion) and a Bronsted acid in Eq. (1.2) (forward direction).
The acidity constants for the above are:
½MOHsurf ½Hþ
K1 ¼ ð1:3Þ
½MOHþ2 surf
and:
½MO surf ½Hþ

K2 ¼ ð1:4Þ
½MOHsurf
At the isoelectric point, ½MO surf ¼ ½MOH2 þ surf , so that Eqs. (1.3) and (1.4)
combine to give:
pK1 þ pK2
pHpzc ¼ ð1:5Þ
2
Introduction 3
where pKi = −log Ki and pHpzc is the isoelectric point. Differences in the surface
acidities of various surface hydroxyls MOH depend on the nature of the metallic
cation M, similar to the way that the identity of an organic radical R influences the
behavior of organic alcohols, ROH.
Nature of Acidic and Basic Surface Sites
A broader view of acids and bases has been given by G. N. Lewis, in which a Lewis
acid is an electron acceptor and a Lewis base is an electron donor [5, 6]. This
definition allows a wide variety of molecules, ions, and compounds to be classified
as Lewis acids or Lewis bases. With this in mind, the following lists the types of
acid sites on metal oxide surfaces. See also Fig. 1.2.
1. unhydroxylated metal ions, Mþds
2. protonated surface hydroxyls, MOH2 þ surf
MOH2 þ surf ! MOHsurf þ Hþ ðaqÞ
3. surface hydroxyls, MOHsurf
MOHsurf ! MO surf þ Hþ ðaqÞ
where ds is the partial charge on a metal surface ion. The first of these acid sites
above is a Lewis acid because the metal ion Mþds can accept electrons [7, 8].
The basic sites [7, 8] on metal oxides are listed below. See also Fig. 1.3.
4. unhydroxylated oxygen anions, Ocs
5. dissociated surface hydroxyls, MO surf
Fig. 1.2 Various types of

acid sites in a metal oxide [8].
+δ refers to a partial charge on
the metal cation and −γ to a
partial charge on the oxygen
ion
Fig. 1.3 Various types of basic sites in a metal oxide [8]. +δ refers to a partial charge on the metal
cation and −γ to a partial charge on the oxygen ion
MO surf þ Hþ ðaqÞ ! MOHsurf
6. surface hydroxyls, MOHsurf
MOHsurf þ Hþ ðaqÞ ! MOH2 þ surf
where cs is the partial charge on an surface oxygen anion. The first of these basic
sites above is a Lewis base because the oxygen anion Ocs can be an electron
donor.
While all these different types of acid or basic sites are possible, the acid-base
properties of the oxide are usually dominated by the behavior of the hydroxyl
groups.
Kaltchev and Tysoe [9] have used NH3 (a Lewis base) to probe surface sites on
thin oxide films on aluminum. Using infrared spectroscopy, these authors found
that NH3 adsorbed on both hydroxylated and dehydroxylated oxide surfaces. In a
study on pure chromium, Ma et al. [10] found that the NH3 molecule reacted with
Cr+3 acid sites in the chromium oxide film formed in vacuum on Cr(100) but did
not react with Cr+3 acid sites in the passive film on chromium. Ma et al. concluded
that the acid-base properties of the passive film on chromium are dominated by the
properties of the OH layer.
Determination of Isoelectric Points of Metal Oxides

and Oxide Films
The isoelectric point of a bulk oxide or oxide film can be measured in several
different ways. Several methods are discussed here. It should be noted that there
have been many more experimental determinations for oxide powders than for
intact oxide films on low-area flat sheets or foils.
Determination of Isoelectric Points of Metal Oxides and Oxide Films 5
Fig. 1.4 The pH dependence

of the zeta potential for the air
formed oxide film on T-6Al-
4V. Redrawn from Ref. [13]
with the permission of
Elsevier
Streaming Potential Measurement
One method of determining the isoelectric point of an oxide-covered metal is to

measure the zeta potential, i.e., the potential drop across the diffuse part of the
electrical double layer, as a function of the pH of the electrolyte. This can be done
by streaming potential measurements [11–15] in which the aqueous phase is forced
to flow across a fixed solid surface. The zeta potential goes through zero at the
isoelectric point of the solid surface, as shown in Fig. 1.4. The basic idea is that the
total charge at the metal/oxide/solution interface, σtotal, is the sum of the charge at
the surface, σsurface, plus that in the diffuse double layer, σdiffuse:
rtotal ¼ rsurface þ rdiffuse ð1:6Þ
At the isoelectric point, σtotal = 0, so that when σdiffuse goes through 0, so does
the surface charge σsurface.
Contact Angle Measurements
Another method is to measure contact angles of aqueous droplets on the oxide-

covered metal surface [16–19]. In this approach, contact angles are measured at the
hexadecane/aqueous solution interface, as show in Fig. 1.5. The pH of the aqueous
phase is varied, and the contact angle is observed to go through a maximum at the
isoelectric point of the oxide film, as shown in Fig. 1.6.
Fig. 1.5 Schematic diagram for a contact angle in the two liquid-solid system [17]
Fig. 1.6 Contact angles as

measured by the goniometer
technique for chromium-
plated steel (Ferrotype) at the
hexadecane/aqueous solution
interface. Redrawn from Ref.
[13] with the permission of
Elsevier
The expression for the surface charge density σ at the oxide/solution interface is:

r ¼ z F CS1
Hþ CS1
O ð1:7Þ
where F is the Faraday, z = 1 here, and Γi refers to surface excesses of H+ or O−.

Use of Eq. (1.7) along with Young’s equation for contact angles, and the Gibbs
expression for surface thermodynamics leads to [16–18]:

d cos h 1 r dw dw
¼ ð2:303 RT + zF Þ þ CS1
O zF ð1:8Þ
d pH c12 zF d pH d pH
At the isoelectric point, the surface charge density σ is zero and the surface con-
centration of dissociated hydroxyl groups CO S1 is also zero, so that Eq. (1.8)
becomes:

d cos h
¼0 ð1:9Þ
d pH r¼0
Fig. 1.7 Schematic diagram showing the interaction of a water molecule with an oxide film
above, at, and below the isoelectric point (IEP) of the oxide. E is the energy of attraction, r is the
distance of approach, and Wadh is the thermodynamic work of adhesion. Reproduced from Ref.
[91] with the permission of Elsevier
At the isoelectric point, the cosine of the contact angle goes though a minimum and
the contact angle through a maximum. Contact angles can be measured directly
using an optical goniometer or indirectly using a Wilhemy balance.
This same conclusion can be reached independently by a qualitative argument
shown in Fig. 1.7. The relationship between the contact angle and the thermody-
namic work of adhesion Wadh is given by the Young-Dupré equation [20]:
Wadh ¼ cL ð1 þ cos hÞ ð1:10Þ
At the isoelectric point of the oxide film, the oxide surface consists of hydroxyl
groups which can interact with water molecules by double hydrogen bonding, as
shown in Fig. 1.7. These interactions are dipole-dipole interactions, for which the
energy of attraction E is [21]:
1
E/ ð1:11Þ
r3
where r is the distance of approach. Above the isoelectric point, water molecules
interact with a negatively charged surface, while below the isoelectric point, water
molecules interact with a positively charged surface, as is also shown in Fig. 1.7.
For these ion-dipole interactions, the energy E is given by [21]:
1
E/ ð1:12Þ
r2
and thus are stronger than interactions at the isoelectric point.

When the energy of attraction is the highest (below and above the isoelectric
point where the stronger ion-dipole interactions occur), then the work of adhesion
Wadh has its highest values. Thus, cos θ also displays its highest values, but θ its
lowest. By the same argument, at the isoelectric point where the oxide surface is
hydroxlyated, the energy of interaction with water molecules is lowest (due to
weaker dipole-dipole interactions). Thus Wadh and cos θ are also lowest, but θ has
its highest value. Thus, the contact angle for aqueous solutions measured as a
function of pH goes though a maximum at the isoelectric point of the oxide surface.
The contact angle method has been used experimentally to determine the iso-
electric point of oxide films on silver [16], chromium, aluminum, and tantalum [17,
18], on an Al-5 % Mg alloy [22], and on aluminum, zinc, and copper [23].
Surface Reaction Method of Simmons and Beard
In this method, developed by Simmons and Beard [24], negatively charged surface
groups MO surf are allowed to react with aqueous K+ ions (a Lewis acid) and
positively charged surface groups MOH2 þ surf are reacted with aqueous anions or
amines (Lewis bases). The extent of the surface reaction is followed by X-ray
photoelectron spectroscopy (XPS) or some other suitable surface analysis
technique.
Above the isoelectric point, the predominant surface species is MO surf , as in
Eq. (1.2). When MO surf , surface groups interact with dissolved K+ ions, then:
MO surf þ Kþ ðaqÞ MO Kþ surf ð1:13Þ
After reaction with K+ ions, the various types of surface sites are given by:
½MOHsurf þ ½MO surf þ ½MO Kþ surf ¼ No ð1:14Þ
The surface concentration of unreacted negatively charged sites MO surf is

negligible for excess K+, so that Eq. (1.14) becomes:
½MOHsurf þ ½MO Kþ surf ¼ No ð1:15Þ

Fig. 1.8 Schematic diagram

showing determination of the
oxide acidity constant pK1
from a plot of the surface
coverage of adsorbed anions
versus the pH of the solution
Use of ½MO surf ¼ ½MO Kþ surf along with Eq. (1.15) in Eq. (1.4) gives:

hþ
pK2 ¼ pH log ð1:16Þ
1 hþ
where hþ ¼ ½MO Kþ surf =No . Equation (1.16) has the form shown in Fig. 1.8. The
surface equilibrium constant K2 is determined from Eq. (1.16) and the pH where
θ+ = 0.5.
Below the isoelectric point, the predominant surface species is MOH2 þ surf , as in
Eq. (1.1). When MOH2 þ surf surface groups interact with dissolved anions A−
(Lewis bases), then:
MOH2 þ surf þ A ðaqÞ MOH2 þ A surf ð1:17Þ
Considerations similar to those for the case of the negatively charged surface
gives the result:

h
pK1 ¼ pH þ log ð1:18Þ
1 h
where h ¼ ½MOHþ 2 A surf =No and is the fractional coverage of reacted

MOH2 þ surf sites. Equation (1.18) has the form shown in Fig. 1.9. The surface
equilibrium constant K2 is determined from Eq. (1.18) and from the pH where
θ− = 0.5.
In this manner, Simmons and Beard determined pK1 = 8.5 and pK2 = 11.5 for
oxide-covered flat surfaces of iron. Thus, the isoelectric point is pHpzc = 10.0 by
Eq. (1.5). This method has also been used by Watts and Gibson [25] who
exchanged surface hydroxyls on the oxide-covered iron surface with either K+
cations or phosphate anions and determined pK1 = 8.4, pK2 = 11.5, and
pHpzc = 10.0, in excellent agreement with the results of Simmons and Beard.
Fig. 1.9 Schematic diagram

showing determination of the
oxide acidity constant pK2
from a plot of the surface
coverage of adsorbed cations
versus the pH of the solution
Ionic Interaction Model of Parks [26]
This model provides a way to correlate the isoelectric points of various oxides (or
oxide films) rather than experimental determination of their values. However, the
isoelectric point of an oxide not yet determined experimentally can be estimated
from this correlation.
The equilibrium between positively and negatively charged surface groups in the
hydroxylated layer is:
MO surf þ 2 Hþ ðaqÞ MOH2 þ surf ð1:19Þ
which combines Eqs. (1.1) and (1.2). The equilibrium constant for Eq. (1.19) is:
½MOH2 þ surf
K¼ ð1:20Þ
½MO surf ½Hþ 2
and the free energy change for Eq. (1.19) is:
DG ¼ 2:303 RT log K ð1:21Þ
At the isoelectric point, ½MO surf ¼ ½MOH2 þ surf and pH = pHpzc. Under these
conditions, Eqs. (1.20) and (1.21) combine to give:

DG ¼ 2:303 2 pHpzc ð1:22Þ
The free energy change can also be calculated from the various surface inter-
actions operative in Fig. 1.10. The surface free energy change consists of contri-
butions from three terms:
Fig. 1.10 Interaction of

protons with an oxide surface
in the ionic interaction model
of Parks [26]

DG ¼ DG ðnon-coulombicÞ þ DG O2 Hþ interactions
ð1:23Þ
þ DG ðMþn Hþ interactionsÞ
The first two terms in Eq. (1.23) may be combined into a constant A′ because
they do not depend on the nature of the metallic cation. Thus,
2 zþ þn
H zM
DG = A0 ð1:24Þ
e½2rðO2 Þ þ rðMþn Þ þ rðHþ Þ
where z is the charge on the proton ðzH þ Þ or on the metallic cation ðzM þn Þ, ε is the
local dielectric constant, and ri are the various ionic radii. With the proton radius
negligible compared to 2r(O−2) + r(Mþn ), Eqs. (1.22) and (1.24) combine to give:
Bzþn
pHpzc ¼ A þ M
ð1:25Þ
R
where A is a constant,
2 zþ
B¼ H
¼ const ð1:26Þ
2:303RT e
and:

R ¼ 2r O2 þ rðMþn Þ ð1:27Þ

Thus, a plot of the isoelectric point versus zðMþn Þ= 2r O2 þ rðMþn Þ is
expected to yield a straight line. Figure 1.11 shows such a plot for various bulk
oxides or oxide films of elemental metals taken from Table 1.1. Ionic radii are from
Shannon and Prewitt [27]. There is considerable scatter in Fig. 1.11, but the
approach allows estimation of unknown isoelectric points. For example, for Li2O,
Fig. 1.11 A Parks-type plot

of the isoelectric point of
various oxides versus the
quantity Z/R, where Z is the
charge on the metal cation of
the oxide and R is twice the
radius of the oxygen anion
plus the radius of the metal
cation in the oxide. Isoelectric
points are taken from
Table 1.1

zðLiþ Þ= 2r O2 þ rðLiþ Þ ¼ 0:282, from which the isoelectric point of Li2O is
calculated to be 13.0.
XPS Correlation Method
Delamar [28] has shown that for metal oxides there is a direct correlation between
XPS shifts and the isoelectric point of the oxide. The idea behind this correlation is
that a lower binding energy for a surface species means that the species is able to
release (i.e., donate) electrons more easily than for a higher binding energy species.
Thus, a lower binding energy corresponds to a better Lewis base [29].
Delamar [28] defined the following quantity involving the XPS O 1s photopeak:
DO ¼ O 1s binding energy 530 ð1:28Þ
Delamar noted that most basic oxides display an O 1s binding energy of

approximately 530 eV, so that ΔO is a measure of the basicity (or acidity) of the
surface oxide. Thus, when ΔO is small, the oxide is basic, and the isoelectric point,
pHpzc, should increase in numerical value.
Delamar also proposed that the acidities (or basicities) of metal cations can be
compared by considering the chemical shift between the zero-valent metal and the
cation. Delamar thus defined a second quantity:
DM ¼ cation binding energy metal binding energy ð1:29Þ
as shown in Fig. 1.12.

Table 1.1 Isoelectric points of some oxides and oxide films

Oxide Isoelectric point References
Oxide film on Ta −0.7 McCafferty and Wightman [17]
IrO2 0.5 Ardizzone and Trasatti [40]
MoO3 1.8–2.1 Compiled by Natishan [41]
SiO2 1.8–2.2 Compiled by Natishan [41]
Ta2O5 2.8 Compiled by Natishan [41]
Oxide film on 304 stainless steel 3–4.8 Boulangé-Petermann [15]
Oxide film on 316 stainless steel 4.2–5.2 Kallay et al. [42]
MnO2 4.7–4.8 Tamura et al. [43]
Oxide film on Ti 2 Kallay et al. [44]
5.0 McCafferty et al. [31, 32]
4.4 Roessler et al. [13]
Oxide film on Ti-6Al-4 V 4.4 Roessler et al. [13]
TiO2 5.0–6.5 Ardizzone and Trasatti [40]
5.5 Hu et al. [33]
SnO2 5–6 Arai et al. [45]
Oxide film on Cr 5.2–5.3 McCafferty and Wightman [17]
CeO2 5.2–6.1 Hsu et al. [46]
6.7 Ozawa and Hattori [47]
ZrO2 5.5–6.3 Compiled by Natishan [41]
5.7–6.1 Mao et al. [48]
Oxidized Zr 6.7 Chvedov and Logan [23]
Cr2O3 6.2–6.3 Compiled by Natishan [41]
NiO 7.3 Moriwaki et al. [49]
8.2–8.6 Kokarev et al. [50]
Fe2O3 7.5 Yates and Healy [51]
Morowaki et al. [49]
8.7 Smith and Salman [52]
Y2O3 8.3 Ozawa and Hattori [47]
CuO 8.5–9.5 Compiled by Kosmulski [53]
Oxidized Cu 9.3 Chvedov and Logan [23]
Al2O3 9.0–9.4 Compiled by Natishan [41]
Oxide film on Al 9.5 McCafferty and Wightman [17]
ZnO 9.2–10.3 Compiled by Natishan [41]
Oxidized Zn 9.3 Chvedov and Logan [23]
Oxide film on iron 9.8 Kurbatov [30]
10.0 Simmons and Beard [24]
CdO 10.3 Janusz [54]
Oxide film on silver 10.4 Chau and Porter [16]
MgO 12.4 Compiled by Natishan [41]
Fig. 1.12 XPS photopeak for

Al 2p3/2 showing the quantity
ΔM. From Ref. [17] by
permission of Elsevier
Fig. 1.13 Isoelectric points

of several bulk oxides or
oxide films as a function of
the quantity (ΔO + ΔM)
Reproduced from Ref. [31,
32] with permission of ECS—
The Electrochemical Society
By arguments given above, small values of ΔM also correspond to Lewis bases.

Delamar showed that there is linear correlation between the sum (ΔO + ΔM) and the
isoelectric point pHpzc of various bulk oxides. See Fig. 1.13. To the data for
Delamar in Fig. 1.13 has been added data for several oxide-covered metals [17, 30].
The data for oxide-covered metals in Fig. 1.13 display some scatter around the data
for bulk metal oxides, but the trend is the same for both bulk oxides and oxide
films.
Thus, Fig. 1.13 can be used to estimate the isoelectric points of various oxide
covered metals from XPS data. As an example, from a previous XPS study on metal
oxide films [4, 31, 32], the quantity (ΔO + ΔM) has been determined to be 5.3 for
oxide-covered titanium. Figure 1.13 gives the isoelectric point of the oxide film on
titanium to be estimated as 5.0. This result is in excellent agreement with the value
of 5.7 for bulk TiO2 determined by Hu et al. [33] from measurements using the
atomic force microscope and with the value of 5.6 for bulk TiO2 determined by
Larson et al. [34] from electrophoretic mobility measurements.
A correlation between the XPS quantity (ΔO + ΔM) and the isoelectric point of
varous oxides has also been observed by Cattania et al. [35].
Fig. 1.14 Schematic

representation of AFM force
measurement between a
negatively charged silica
sphere and a TiO2 substrate in
an aqueous solution. Redrawn
from Hu et al. [33] with the
permission of the American
Chemical Society
Measurement of Near-Surface Forces
A more recent method of determining the isoelectric point of an oxide or oxide film
involves measurements with an atomic force microscope (AFM). Hu et al. [33]
measured the force between a spherical probe placed at the end of an AFM can-
tilever and a flat TiO2 substrate, as shown in Fig. 1.14. The charge within the
diffuse double layer is calculated from the force between the electrode and the tip of
the AFM cantilever. At the isoelectric point, the charge on the diffuse double layer
is zero, and the net surface charge is zero. Similar studies have been carried out on
TiO2 [36], SiO2 [37], Al2O3 [38], and oxide-covered tungsten [39].
Numerical Values of Isoelectric Points
Table 1.1 lists the isoelectric points for various bulk oxides and oxide films. The
isoelectric point is a function of the identity of the individual cation in the M–OH
bond. Oxides with an isoelectric point less than 7 are acidic oxides, and oxides with
an isoelectric point greater than are basic oxides. It can be seen that there is a wide
range in the values for oxide-covered metals. For instance, tantalum which is a
passive metal has an oxide which is acidic, and aluminum, which is used in many
applications in neutral solutions has an oxide which is basic. Much more data are
available on bulk oxides than on intact oxide films, but in general, the isoelectric
point of an oxide film is similar to that for the stand-alone bulk oxide.
Fig. 1.15 Surface charge

character of several oxide
films. “IEP” refers to the
isoelectric point of the oxide
film
Surface Charge of Oxide Films
Figure 1.15 illustrates the surface charge character of several different oxides. As
seen in Table 1.1, the isoelectric point of an oxide-covered aluminum surface is 9.5.
Thus, in a solution of pH 7, the surface of aluminum oxide (and oxide-covered
aluminum) consists of acidic groups (positively charged ¼ AlOH2 þ ). That is, in an
aqueous solution of pH 7, the oxide film has a positively charged surface, as in
Fig. 1.15 and Eq. (1.1). If the pH of the aqueous solution is greater than 9.5, then
the oxide surface will be negatively charged, as per Fig. 1.15 and Eq. (1.2).
For some oxides in solutions of pH 7, such as tantalum oxide, the surface
consists predominantly of basic groups (negatively charged TaO−), as in Fig. 1.15.
Oxide films on tantalum will have a positively charged surface only in acidic
solutions where the pH is below the value of the isoelectric point.
Pitting of Aluminum
Most passive metals are subject to pitting corrosion. Pitting is a highly localized
breakdown of the passive film caused by the presence of an “aggressive” anion in the
electrolyte, usually Cl− ions. However, pitting of various meals or alloys has also
occurred in the presence of other anions, including Br−, I−, SO4 2 , or NO3 [55].
The chloride ion has a special importance in pitting corrosion for several reasons.
First of all, Cl− ions are abundant, being constituents of seawater, brackish waters,
de-icing salts, and of airborne salts. In addition, chlorides are found in the human
body so they can cause the pitting of biomedical implants. Chlorides are also
common contaminants in various electronic systems due to handling and processing.
Pitting of Aluminum 17
The chloride ion is a strong Lewis base (electron donor) and tends to interact with
Lewis acids (electron acceptors), such as metal cations. In addition, the chloride ion is
a relatively small anion and has a high diffusivity.
Adsorption of Cl− on Aluminum
The first step in the pitting process is the adsorption of chloride on the oxide-
covered surface.
For an oxide-covered surface of aluminum immersed in a neutral solution, the
aluminum oxide surface will be positively charged, as discussed above. This sit-
uation will favor the adsorption of the negatively charged Cl− ions onto the posi-
tively charged oxide surface comprised predominantly of ¼ AlOH2 þ groups. The
attractive forces which result when an anion, such as chloride, interacts with an
ionic surface, such as an oxide, consist of (i) coulombic forces, (ii) induction of the
adsorbent by the approaching ion, (iii) electrostatic polarization of the ion, and (iv)
non-polar van der Waals forces [56].
Of these attractive forces, the largest are the first two interactions, which are
ionic, as shown in Table 1.2. Thus, adsorption of chloride ions is most favored on a
positively charged oxide surface for which the ion-ion forces are attractive in
nature. When the oxide surface is negatively charged, ion-ion forces are not
attractive in nature so that adsorption of Cl− onto the negatively charged oxide
surface is much less favored, but can proceed through the operation of van der
Waals (dispersion) forces.
The above trends are also consistent with acid-base considerations. The nega-
tively charged Cl− anion is a Lewis base (electron donor) and the ¼ AlOH2 þ
surface group is a Lewis acid (also a Bronsted acid), so interaction between these
two reactants is favored. However, adsorption of the Lewis base Cl− would not be
favored on the Lewis base surface species, = AlO−, which is the predominant
surface species for pH values above the isoelectric point.
XPS studies have confirmed that the amount of chloride taken up by the passive
film on aluminum varies with solution pH. Kolics et al. [57] observed that the
minimum concentration of surface chloride occurred at a pH of about 9.5, i.e., at the
isoelectric point, where the surface has a net charge of zero. Moreover, the amount
Table 1.2 Forces involved in the interaction of an anion with an oxide surface. From Ref. [56]
Type of interaction Distance dependence of energy of interaction
Coulombic 1/r
Electrostatic induction of 1/r4
substrate by ion
van der Waals (dispersion) 1/r6
Electrostatic polarization Usually small. Calculated from the polarizability of the ion and
the ion the electrostatic field of the adsorbent
Fig. 1.16 Chloride content in

the oxide film on aluminum as
a function of pH after
immersion in a 0.1 M solution
of NaCl. Redrawn from
Kolics et al. [57] with the
permission of ECS—The
Electrochemical Society
of adsorbed chloride increased with decreasing pH, in agreement with the above
discussion. See Fig. 1.16. A similar observation has been reported for the
adsorption of Cl− on bulk aluminum oxide [58].
Electrode Kinetics of Pit Initiation on Aluminum
The first step in the pitting process of any metal in chloride-containing solutions is
the adsorption of Cl− ions on the oxide-covered metal surface.
In the case of aluminum, the isoelectric point of the oxide-covered aluminum
surface is 9.5. Thus, when aluminum is used in neutral or nearly neutral solutions
(the usual case), there is a positive charge on the oxide surface. This situation is
favorable for adsorption of Cl− ions on the oxide film on aluminum, as has been
discussed above. Subsequent steps in the pitting process involve transport of
chloride through the oxide film and local dissolution of aluminum atoms at the
metal/oxide interface. See Fig. 1.17. Experimental evidence supporting this
sequence of events has been given elsewhere [59], and more detail on the inter-
action of Cl− ions with passive films is discussed later in this volume.
The following sequence of reactions [60] leading to pit initiation takes into
account adsorption of chloride ions, transport of chloride ions through the oxide
film by means of oxygen vacancies, and local dissolution of aluminum atoms at the
metal/oxide interface in three consecutive one-electron transfer reactions:
Fig. 1.17 Sequence of steps in the pitting of aluminum by chloride ions
1
AlðoxideÞOH þ Hþ AlðoxideÞOH2 þ ð1:30Þ
1
2
AlðoxideÞOH2 þ þ nCl AlðoxideÞOH2 þ Cln n ð1:31Þ
2
3
AlðoxideÞOH2 þ Cln n þ nVO Al½ðnClO ÞðoxideÞOH2 þ ð1:32Þ
3
4
Al½ðnClO ÞðoxideÞOH2 þ Alþ ½ðnClO ÞðoxideÞOH2 þ þ e ð1:33Þ
4
5
Alþ ½ðnClO ÞðoxideÞOH2 þ Alþþ ½ðnClO ÞðoxideÞOH2 þ þ e ð1:34Þ
5
6
Alþþ ½ðnClO ÞðoxideÞOH2 þ ! Alþþþ ½ðmClO ÞðoxideÞOH2 þ þðn mÞCl þ e
ðat site of pitÞ
ð1:35Þ
In the notation Al(oxide)OH, Al refers to a substrate aluminum metal atom

located immediately beneath the oxide film. “Oxide” refers to the oxide film cov-
ering that aluminum atom, Al, which will eventually undergo localized dissolution,
i.e., pitting. Finally, in the notation Al(oxide)OH, OH refers to the outer layer of
surface hydroxyl groups. In addition, VO refers to an oxygen vacancy in the film,
ClO: is a chloride ion occupying an oxygen lattice site, and Al½ðnClO: ÞðoxideÞOH2 þ
refers to the oxide film on aluminum through which chloride ion transport is
occurring. The symbol Al½ðnClO: ÞðoxideÞOH2 þ represents zero valent aluminum
metal existing prior to pit initiation in which the metal substrate is covered by its
chloride-containing oxide film. Alþ ½ðnClO: ÞðoxideÞOH2 þ ; Alþþ ½ðnClO: ÞðoxideÞ

OH2 þ , and Alþþþ ½ðmClO: ÞðoxideÞOH2 þ respectively, refer to mono-, di-, and tri-
valent aluminum ions at the metal/oxide interface, i.e., at the site of the initiating pit.
As also discussed elsewhere [60], Eq. (1.31) allows that chloride clusters form
on the oxide-covered surface and that the concentration of chloride ions in the oxide
film decreases with the onset of pitting [61]. Thus, 0 < m < n in Eq. (1.35) allows
that there is some expulsion of chloride from the film in the vicinity of the pit when
pitting initiates. The electrolyte is provided by the presence of water in the passive
film [62, 63].
The steady state solution for the anodic current density ia due to Eqs. (1.30)
through (1.35) is [60]:

ia k2 k3 k4 k5 k6 N0 ½VO n ½Hþ
¼3 ½CI n e5FE=2RT ð1:36Þ
F k2 k3 k4 k5 ½Hþ þ K1
where F is the Faraday, ki is the rate constant for the ith step, N0 is the total number of
surface sites (in moles/cm2), ½VO is the concentration of oxygen vacancies in the
film, k1 =k1 ¼ K1 , the surface equilibrium constant in Eq. (1.30), and E is the
electrode potential. The electrochemical transfer coefficient has been taken to be 1/2.
This expression, i.e., Equation (1.36), can be used to explain (i) the variation in
pitting potential with pH (at fixed Cl− concentration) and (ii) the variation in pitting
potential with Cl− concentration (at fixed pH).
Pitting Potential of Aluminum as a Function of PH
The pitting potential of aluminum is independent of pH for pH values between 4 and

8 [60, 64], as shown in Fig. 1.18. Over this pH range, the oxide-covered aluminum
surface consists of ¼ AlOH2 þ groups, as discussed above, adsorption of Cl− is
favored, and pitting proceeds at the same critical electrode potential for each pH.
As also shown in Fig. 1.18, the pitting potential increases with pH for pH values
at and above the isoelectric point (9.5). This is because Cl− adsorption is less
favored on either = AlOH, or = AlO− surface groups so that it is necessary to move
the electrode potential to more positive values to first adsorb chloride ions and then
to drive them into interior of the passive film [57, 59, 61, 65].
The experimental results shown in Fig. 1.18 can be explained by application of
Eq. (1.36). Any small value of ia above the background value for the passive film
current (say ia ¼ icrit ) can be chosen to consistently mark the start of pitting. Thus,
when ia ¼ icrit ; E ¼ Epit , the critical pitting potential.
For pH values well below the isoelectric point, pH \ pK1 ; ½Hþ [ K1 , so that
½H =ð½Hþ þ K1 Þ ffi 1. Taking logarithms in Eq. (1.36) under these conditions and
þ
differentiating with respect to the pH at constant chloride concentration gives:

Fig. 1.18 Pitting potential of

aluminum as a function of pH
at two different fixed
concentrations of chloride
ion. From Ref. [60] by

d Epit
ffi0 ð1:37Þ
d pH ½CI
in agreement with experiment.

For pH values above the isoelectric point, pH [ pK1 ; [Hþ \ K1 , so that
½H =ð½Hþ þ K1 Þ ffi ð½Hþ =K1 . Taking logarithms in Eq. (1.36) under these con-
þ
ditions and again differentiating with respect to the pH at constant chloride con-
centration now gives:

d Epit 2 2:303 RT
¼ ¼ const ð1:38Þ
d pH ½CI 5 F
Thus, above the isoelectric point, the pitting potential increases linearly with pH
(at constant chloride ion concentration). Figure 1.18 shows that this predicted
relationship holds, although Fig. 1.18 contains limited data points above the iso-
electric point (due to the general corrosion of aluminum at higher pH values. The
predicted value of dEpit =dpH from Eq. (1.38) is 0.024 V, while experimental values
from Fig. 1.18 are 0.053 V for 0.1 M Cl− and 0.025 V for 1 M Cl−.
Pitting Potential of Aluminum as a Function of Cl−

Concentration
Figure 1.19 shows experimentally determined pitting potentials for pure aluminum
as a function of chloride ion activity (concentration) for chloride concentrations of
0.01–1.0 M [60]. It can be seen that the pitting potential varies linearly with the
logarithm of the chloride concentration at constant pH.
Fig. 1.19 Pitting potential of

aluminum as a function of
chloride ion concentrations at
a fixed pH of 7. From Ref.
[60] by permission of Elsevier
This behavior also follows from Eq. (1.36). Taking logarithms with respect to
chloride ion concentration at constant pH gives:

d Epit 2 2:303 RT
¼ n ¼ const0 ð1:39Þ
d log ½CI pH 5 F
The parameter n was evaluated from the variation in the steady-state anodic
current at the pitting potential as a function of chloride ion concentration [60], with
the result being n = 3.77. This average number of chloride ions n = 3.77 associated
with a pitting site is within the range of values (2–8) as determined from induction
times to pitting for aluminum and its alloys in various aqueous solutions [66, 67].
With n = 3.77, Eq. (1.39) becomes:

d Epit
¼ 0:0891 ð1:40Þ
d log ½CI pH
in nominal agreement with the experimentally determined value of −0.0779, as

shown in Fig. 1.19.
Experimental values for the variation in pitting potential of pure aluminum with
the logarithm of the chloride concentration have been reported to range from
−0.073 to −0.12 V [68–72], although Bockris and Minevski [73] report a much
higher value of ca. −0.315 V. Equation (1.40) thus agrees with the experimental
observation that the pitting potential varies linearly with the logarithm of the
chloride concentration at constant pH. Variations in experimental values for
dEpit =dlog[Cl reported in the literature may be due to variations from system to
system in the number n of chloride ions associated with a pit initiation site.
Thus, the dependence of the pitting potential on the chloride ion concentration or
on the pH is described by the electrode kinetic mechanism considered here.
Acid-Base Interactions in Metal/Polymer Adhesion 23
Acid-Base Interactions in Metal/Polymer Adhesion
Introduction
The adhesion properties of an organic coating and its corrosion performance are
related. If corrosion occurs beneath the organic coating, a loss of adhesion of the
coating will ensue. On the other hand, if the coating does not possess good adhesion
to the metal substrate, then undermining of the coating will produce localized
pockets of electrolyte which promote corrosion.
The relationship between adhesion and corrosion is seen schematically in
Fig. 1.20.
Theories of Polymer Adhesion
There are several fundamental theories of metal/polymer adhesion. These are [74]:
1. The adsorption theory (physical and chemical adsorption), in which the adhe-
sive bond depends on adsorption forces between the two interfaces being joined,
2. The chemical reaction theory, in which a chemical reaction occurs between the
polymer and the metal,
3. The mechanical interlocking theory, in which adhesion is due to the mechanical
interlocking or keying of the polymer into cavities and pores of the metal
surface,
Fig. 1.20 Relationship between adhesion of an organic coating and corrosion of the metal
substrate
4. The electrostatic theory, in which adhesion depends on the existence of an

electrostatic charge between the two surfaces.
Each of these last three mechanisms can operate under certain conditions, but the
most pervasive mechanism is the first, in which the polymer adheres to the metal
substrate (or, more properly, to its oxide film) because of interatomic and inter-
molecular bonds which are formed. These adsorption forces that exist across all
bonding interfaces include Lewis acid/Lewis base forces.
All metal surfaces to be coated with an organic polymer contain either the
incipient air-formed oxide film or an oxide produced by a surface treatment (such as
chromates or their replacements). Thus, it is appropriate to consider the interaction
of organic polymers with oxide films.
Acid-Base Nature of Organic Polymers
As with oxide films, organic polymers can be also be classified as Lewis bases
(electron donors) or Lewis acids (electron acceptors). For example, poly (methyl
methacrylate) (PMMA) is a typical basic polymer, as can be reasoned from its
molecular structure, as shown in Fig. 1.21. The carbonyl oxygens of esters are basic
(electron-donating) sites, which can form acid/base bonds with electron-accepting
sites of Lewis acids [75].
Poly(vinyl chloride) (PVC) is an example of a polymer which is a Lewis acid.
For PVC, it is the partially positive -CH portion of the vinyl group which is the
electron-accepting site and which can form acid/base bonds with electron-donating
sites of Lewis bases [75]. See Fig. 1.22.
Fowkes and Mostafa [76] found that PMMA adsorbed strongly onto silica (an
acidic oxide) but not onto calcium carbonate (a basic solid), whereas chlorinated-
PVC (an acidic polymer) adsorbed onto calcium carbonate but not onto silica. Thus,
Fowkes [77] reasoned that basic polymers adsorb only onto acidic solids and acidic
polymers adsorb only onto basic solids. To carry this a step further, basic polymers
should adsorb onto acidic oxide films and acidic polymers onto basic oxide films. In
addition, the practical adhesion of basic polymers onto metals is expected to be the
greatest on acidic oxide films. Similarly, the practical adhesion of acidic polymers
Fig. 1.21 An example of a basic polymer

Fig. 1.22 An example of an

acidic polymer
onto metals should be the greatest on basic oxide films. This is because a bond is
formed in which electrons donated by the Lewis base are shared with the Lewis
acid site.
Characterization of the Acid-Base Properties of Polymers
The acid-base properties of polymers can be determined by several techniques,

including streaming potential measurements [78] (as in the case of oxide-covered
metals), inverse gas chromatography [78, 79], and the contact angle method of van
Oss and Good [80–83].
Inverse gas chromatography is a chromatographic technique in which the solid
phase is of main interest rather than the gas phase, and hence the designation
“inverse”. This technique consists of measuring the retention times of various gases
or vapors on a column packed with powders, fibers, or films of the polymer of
interest. By using probing gases or vapors of known acid-base character, the acid-
base nature of the polymer can be determined.
The contact angle method of van Oss and Good [80–83] uses various wetting
liquids of various acid-base properties to probe the acid-base nature of the polymer
to which they are applied.
This method extends the work of Fowkes [76], who wrote the surface free
energy of a pure phase as the sum of several contributions:
c ¼ c d þ ci þ cp þ ch þ ð1:41Þ
where cd = van der Waals dispersion forces, ci = Debye induction forces (dipole-
induced dipole), cp = Keesom dipole-dipole forces, and ch = the contribution due to
hydrogen bonding. Equation (1.41) was rewritten by van Oss and Good [80–83] as:
c ¼ cLW þcAB ð1:42Þ
where cLW ¼ cd þ ci þ cp are called Lifshitz-van der Waals forces, and cAB are
acid-base forces due to hydrogen bonding and all electron donor-electron acceptor
interactions. By using various combining rules across interfaces, van Oss and Good
derived the expression:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
þ þ
ð1 þ cos hÞcL ¼ 2 ðcLWS c LW
L Þ þ 2 ðc S L c Þ þ 2 ðc S cL Þ ð1:43Þ
where θ is the contact angle of a liquid on a solid surface, cL is the surface tension
of the spreading liquid, and cL LW ; cL þ , and cL are the Lifshitz-van der Waals,
acid, and base contributions, respectively, to the total surface tension, cL . Similarly,
cS LW ; cS þ , and cS are the Lifshitz-van der Waals, acid, and base contributions,
respectively, to the total surface free energy of the solid, cS . The quantities
cL ; cL LW ; cL þ , and cL are known for a test set of standard wetting liquids. The
terms cS LW ; cS þ , and cS are the unknown quantities. Equation (1.43) is an
equation in three unknowns: cS LW ; cS þ , and cS , as can be seen more clearly by
rewriting Eq. (1.43) in matrix form:
2 12 12 þ 12 32 LW 12 3 2 3
cLW c cL1 c
6 L1 L1
þ 12 7 6 S 1 7 1 cL1 ð1 þ cos h1 Þ
6 cLW 12
c
12
cL2 7 6 þ 2 7 ¼ 4 cL2 ð1 þ cos h2 Þ 5 ð1:44Þ
4 L2 L2 5 4 cS 5 2
LW 12 12 þ 12 12 cL3 ð1 þ cos h3 Þ
cL3 c
L3 cL3 cS
where the numerical subscripts refer to the wetting liquid. Equation (1.44) can be
written as:
AX = b ð1:45Þ
where A is the matrix containing the various surface tension parameters.

The first of the unknowns in Eq. (1.44), cS LW , is determined by using a nonpolar
liquid, such as methylene iodide, for which both cS þ and cS are zero and
cL LW ¼ cL . Under these conditions, Eq. (1.44) reduces to:
ð1 þ cos hÞ2
S ¼ cL
cLW ð1:46Þ
4
which allows evaluation of the parameter cS LW from the surface tension cL of

methylene iodide and from the contact angle h.
After this step, Eq. (1.44) then contains two unknowns cS þ and cS . These
unknowns are determined by using two different polar liquids and solving the
resulting two equations simultaneously. One of the polar wetting liquids used is
water and is taken pairwise in turn with formamide, ethylene glycol, or glycerol.
These combinations of wetting liquids (plus methylene iodide) gave the best con-
ditioned matrices A in Eq. (1.45). A matrix A is ill-conditioned if relatively small
changes in the entries of A can cause relatively large changes in the solutions to
Ax = b. The smaller the condition number, the more well-conditioned is the matrix.
More detail has been given elsewhere [18].
The method of van Oss and Good has been used to determine the acid/base
nature of the following polymers: poly(vinylchloride), poly(methylmethacrylate),
poly(vinylacetate), and a commercial pressure sensitive adhesive (adhesive side).
Table 1.3 lists the contact angle data and the resulting values of cS LW , cS þ , and cS
for the various polymers.
For any given polymer, there is some variation in the values of cS þ , andcS . As
noted by Good and van Oss [82, 83], none of the solid surfaces display high values
of cS þ , but the acidic polymer surfaces have higher values of cS þ than do the basic
polymer surfaces. The basic surfaces clearly have higher values of cS than do the
acidic surfaces.
Comparison of the values of cS þ , and cS for the pressure sensitive adhesive
with those for the acidic and basic polymers shows that pressure sensitive adhesive
has an acidic surface.
A compilation of some acidic and basic polymers is given in Table 1.4.
Table 1.3 Surface free energy components (in mJ/m2) for various polymers as determined by the
method of van Oss and Good [18]
Water plus
Polymer Wetting liquid θ cS LW cS þ cS
Poly(vinyl chloride) Methylene iodide 38.8 ± 2.6 40.2 – –
(Acidic) Water 87.8 ± 4.2 – –
Formamide 69.2 ± 2.9 0.33 4.8
Ethylene gylcol 64.7 ± 2.6 0.15 4.1
Glycerol 84.2 ± 0.9 0.78 6.3
0.42 5.1
Poly(methyl methacrylate) Methylene iodide 23.9 ± 0.8 46.5 – –
(Basic) Water 64.3 ± 1.8 – –
Formamide 47.0 ± 1.8 0.00040 16.1
Ethylene gylcol 47.1 ± 0.8 0.13 19.3
Glycerol 64.3 ± 2.2 0.11 19.0
0.080 18.1
Poly(vinyl acetate) Methylene iodide 33.8 ± 3.6 42.6 – –
(Basic) Water 60.6 ± 2.9 – –
Formamide 46.2 ± 3.8 0.064 19.9
Ethylene gylcol 47.1 ± 1.9 0.059 24.6
Glycerol 61.6 ± 2.9 0.00082 22.5
0.041 22.3
Pressure sensitive Methylene iodide 90.2 ± 3.9 12.6 – –
adhesive (Acidic) Water 97.7 ± 5.3 – –
Formamide 97.6 ± 4.8 0.096 10.7
Ethylene gylcol 82.3 ± 5.7 0.53 5.0
Glycerol 108.0 ± 6.2 0.63 14.1
0.42 9.9
Table 1.4 Some examples of

Acidic polymers
acid, basic, or amphoteric
polymers Poly(vinyl chloride) [75]
Poly(vinyl fluoride) [75]
Poly(vinylidene fluoride) [75, 82]
Poly(vinyl butyral) [75]
Chlorosulfonated polyethylene [84]
Pressure sensitive adhesive composed of poly (acrylate-co-
acrylic acid) [18, 85]
Polypropylene treated in an O2 plasma [86]
Polycarboxylic acids [87]
Basic polymers
Poly (methyl methacrylate) [75, 82]
Poly (vinyl acetate) [18]
Polyethylene oxide [82]
Polyisoprene [84]
Polybutadiene [84]
Poly (styrene) [88]
Cellulose acetate [82]
Polypropylene treated in a N2 plasma [86]
Poly (2-vinylpyridine) [87]
Amphoteric polymers
Rubber surfaces [84]
Poly (acrylic acid) [89]
Adhesive protein of common blue mussel [90]
Metal/Polymer Adhesion Tests
There have been but a few studies which show clearly the relationship between
metal/polymer adhesion and the acid-base properties of the system. An experi-
mental study by the author which involves two of the polymers characterized above
is be discussed here [18, 91].
The adhesion of the pressure-sensitive adhesive (PSA) to various oxide-covered
metal substrates was determined by measuring the peel force, as shown in Fig. 1.23.
To minimize viscoelastic processes and to compare the interfacial adhesion for the
various metal substrates, a moderately slow peel rate of 1.2 cm/min was used. The
peel force was measured for a 1.91 cm (3/4 inch) wide strip of the pressure-sensitive
adhesive on aluminum, chromium, tantalum, and zirconium for various contact
times before the peel measurement. In each case, the metal sample was first cleaned
in an argon plasma prior to application of the pressure-sensitive adhesive, which
was rolled in place using a 5 pound (2.27 kg) cylindrical roller.
Figure 1.24 shows the peel force measured at 27 °C for the pressure-sensitive
adhesive on various metal surfaces as a function of the surface isoelectric point of
Fig. 1.23 Schematic diagram showing a 180° peel test for polymer adhesion
Fig. 1.24 Measured peel force for a pressure-sensitive adhesive (an acidic polymer) versus the
isoelectric point of the oxide film. The peel force is shown for various contact times prior to the
peel measurement [3, 18]. Reproduced by permission of Taylor and Francis
the oxide film. Figure 1.24 shows that the peel force for the acidic PSA increases
with increasing basicity of the oxide film and is clearly the highest for the most
basic oxide film, i.e., the oxide film on aluminum. This means that there is a greater
ability for this basic oxide film to donate electrons to the acidic polymer. This
increased acid-base interaction leads in turn to an increased practical adhesion.
The adhesion of a basic polymer, poly(methyl methacrylate), was measured by
measuring the pull-off force from oxide-covered substrates of silicon, chromium,
and aluminum. Experimental details have been given elsewhere [91], but in brief,
metal samples were first solvent cleaned, films of poly(methyl methacrylate) were
spin cast from chlorobenzene solutions to a thickness of nominally 1 μm in
thickness, and 7.2 mm diameter aluminum studs were attached to the surface of the
poly(methyl methacrylate) film using a commercial epoxy which was allowed to
Fig. 1.25 Diagram of a pull-off test specimen showing an aluminum stud attached to a poly
(methyl methacrylate) coating on an oxide-covered metal substrate [91]
cure for 2–3 days. Pull-off measurements were then made using a Romulus III stud
pull tester at a loading rate of 1.5 N/s. The tester pulls the stud normal to the surface
until the poly(methyl methacrylate) film is removed from the metal substrate at a
critical force. See Fig. 1.25.
Figure 1.26 shows the measured pull-off force per unit area for the basic polymer
PMMA for three different oxide-covered metal surfaces.
The measured pull-off force includes both interfacial and viscoelastic contribu-
tions, but because the polymer is the same for each metal surface, the viscoelastic
contributions may be taken to be the same in each case. Thus, differences in the
measured pull-off force may be ascribed to differences in interfacial adhesion. As
seen in Fig. 1.26, the pull-off force for the basic polymer PMMA is highest for the
most acidic oxide film. This means that there is a greater ability for this acidic oxide
film to accept electrons from the electron-donating basic polymer. This increased
acid-base interaction leads in turn to an increased practical adhesion.
In other related work, Chvedov and Logan [23] observed that the adhesion
strength of two basic polymers on commercial grade aluminum decreased with the
isoelectric point (i.e., with increasing basicity of the aluminum surfaces. The alu-
minum was given chromate-free pretreatments, which varied the isoelectric point.
The adhesion strength was measured using a pull-tab test.
Fig. 1.26 Measured pull-off

force for poly(methyl
methacrylate (a basic
polymer) versus the
isoelectric point of the oxide
film [3, 91]. Reproduced by
permission of Taylor and
Francis
Starostina et al. [92] found that the adhesion of an amine-modified polystyrene

polymer increased with the relative acidity of the underlying oxide film on the metal
substrate. The extent of adhesion was measured from cathodic delamination
experiments in which the diameter of the delaminated defect was measured.
Thus, experimental results with either basic or acidic polymers on oxidized
metals having various acid-base properties provide direct experimental evidence as
to the importance of Lewis acid-Lewis base effects in the adhesion of polymers to
oxide-covered metal surfaces.
Wet Adhesion
The situation becomes more complicated when water molecules accumulate

beneath the organic coating because water interferes with the adhesion between the
organic coating and the metal substrate. The extent of adhesion in the presence of a
liquid (usually water) is referred to as wet adhesion and is considered by Funke [93]
to be the most important property of the organic coating. Two different reasons have
been proposed for the loss of adhesion in the presence of water [94, 95]. These are:
1. Chemical disbonding due to the interaction of water molecules with covalent,
hydrogen, or polar bonds between the organic polymer and the oxide film,
2. Mechanical disbonding due to forces caused by accumulation of water and
osmotic pressure.
At present the mechanisms of wet adhesion (and its loss) are not completely
understood.
Acid-Base Effects in Corrosion Inhibition
In keeping with the theme of this chapter, corrosion inhibition will be considered
for nearly neutral or basic aqueous solutions, in which the metal surface is covered
with an oxide film. Acid solutions will not be considered because in acid solutions
the metal surface generally does not contain an oxide film; and the inhibitor
interacts with the metal surface without the intervening effect of the oxide film.
General Effect of Oxide Films
It is well known that organic inhibitors can adsorb strongly onto bare metal (i.e.,
oxide-free) surfaces, but corrosion inhibitors can also interact strongly with oxide
films. Table 1.5 compares the heats of adsorption DHads for several small organic
molecules onto the surfaces of metals, oxides, or oxide films. The heats of
Table 1.5 Heats of adsorption from the vapor phase at near-zero coverages for various small
organic molecules or water on metals, oxides, or oxide-covered metal surfaces
Adsorbate ΔHads on metal (kJ/mole) ΔHads on metal oxide or References
oxide film (kJ/mole)
n-C3H7NH2 −226 (on nickel) −84 (on oxidized nickel) Yu et al. [96]
−105 (on copper) −96 (on oxidized copper)
n-C3H7OH −347 (on nickel) −138 (on oxidized nickel) Yu et al. [96]
−105 (on copper) −67 (on oxidized copper)
– −200 (on Al2O3) Rossi et al. [97]
n-C3H7NH2 −138 (on iron) −188 (on oxidized iron) Yu Yao [98]
n-C4H9NH2 −144 (on iron) −186 (on oxidized iron)
(n-C3H7)2NH −128 (on iron) −183 (on oxidized iron)
(C2H5)3N −156 (on iron) −207 (on oxidized iron)
/-CH3 – −69 to −77 (on ZnO) Nakazawa [99]
−83 (on FeO)
ϕ-OH* – −131 (on ZnO Nakazawa [99]
−140 (on FeO)
H2O −113 (on oxidized iron) Yu Yao [98, 100]
−134 (on α-Fe2O3)
−134 (on α-Al2O3)
−80 to −90 (on NiO) Matsuda [101]
D2O −89 (on ZnO) Nakazawa [99]
−65 (on FeO)
*The symbol ϕ denotes the benzene ring
adsorption are from the vapor phase, so as to consider the energy of interaction
between the adsorbate and the surface without the complicating effect of the sol-
vent. The heats of adsorption in Table 1.5 are for near-zero coverages, so that
values of DHads are those which pertain before DHads begins to decrease with
increasing coverage. It can be seen that in some cases, the heats of adsorption onto
the oxide or oxide-covered surface are well within the range expected for a
chemisorptive process. This strong interaction is due to interaction of dipoles in the
organic molecule with cations or anions present in the oxide surface. In addition,
π-bonding is also possible between benzene rings and the metal oxide.
Because the metal surface is oxide-covered, the role of the inhibitor in neutral
solutions is to make oxide films protective and to keep them so. Oxidizing inhib-
itors, like chromates, react with the oxide-covered surface and are incorporated into
the passive film. With non-oxidizing inhibitors, however, such chemical reactions
do not occur. Phosphates, for example, are incorporated into the passive film by
exchange with surface oxide or hydroxide ions [102]. Other inhibitors, such as
Acid-Base Effects in Corrosion Inhibition 33
sodium azelate are not taken up as uniformly, but instead function by repairing
pores in oxide films by reacting there with dissolved cations to form insoluble local
precipitates [103].
Mechanisms of Interaction with Oxide Films
There are two predominant mechanisms by which inhibitors can interact with oxide
films. Both of these mechanisms involve Lewis acid-Lewis base effects:
1. Inhibitor molecules can react with hydroxylated species at the oxide surface, i.e.,
with MOHsurf at the isoelectric point, with MOH2 þ surf below the isoelectric
point, or with MO surf above the isoelectric point.
2. The inhibitor can interact with metal cations Mþn located in the surface of the
oxide film to form surface chelates, with the metal cation being the central ion of
the chelate. Each cases is considered below.
Interaction of Inhibitors with Hydroxylated Oxide Films
The effect of the surface charge of the oxide film on corrosion inhibition can be
illustrated by various examples.
The corrosion rate of zinc pigments in aqueous solutions containing phenolates
inhibitors is higher at pH 10 than at pH 8 [104]. The authors attribute this difference
to surface charge effects related to the isoelectric point of the oxide film on zinc
surfaces. At pH values less than the isoelectric point of ZnO (pHpzc 9), the ZnO
surface is positively charged so that more negatively charged phenolates can be
adsorbed, lowering the corrosion rate. Above the isoelectric point of ZnO, the ZnO
surface is negatively charged, and adsorption of phenolate ions is not favored.
Similar results have been obtained [105] for the inhibition of aluminum by poly-
mers with carboxyl groups, which are Lewis bases, i.e., electron donors. This effect
is illustrated schematically in Fig. 1.27.
The interaction of oxide films with an aggressive ion, i.e., Cl , as a function of
pH has been considered previously in this chapter.
Chelating Compounds as Corrosion Inhibitors
Chelating compounds are organic molecules with at least two polar functional
groups capable of ring closure with a metal cation. The functional groups may
either be basic groups, such as NH2 , which can form bonds by electron donation,
Fig. 1.27 Schematic representation of the adsorption of a negatively charged corrosion inhibitor
as a function of the surface charge character of an oxide film
or acidic groups, such as −COOH, which can coordinate after the loss of a proton.
An example of a surface chelate is shown in Fig. 1.28. The chelating molecule may
interact with metal ions which exist in the oxide film, or they react with metal
cations which are first produced by metallic dissolution. In the latter case, a com-
plex of high molecular weight and low solubility is precipitated near the metal
surface, and a barrier film is formed. Other chelating agents include mercaptoacetic
acids, ethyene diaminetetraacetic acid, and catechols. See Fig. 1.29. In each case,
surface chelates are formed involving bonding between a surface cation and an
oxygen, sulfur, or nitrogen atom in the chelating molecule.
It is well known that in bulk solutions, 5-membered chelate rings are the most
stable [106], and it appears that this effect carries over to surface chelates as well.
There are steric requirements for surface chelates so that not all compounds which
are capable of forming chelates in the bulk solution are effective corrosion
inhibitors.
Chelating compounds have been studied as corrosion inhibitors for steel in
industrial cooling waters [107, 108], and for zinc [109, 110] and aluminum [111, 112]
Fig. 1.28 A surface chelate

formed between 1, 10
phenanthroline and the oxide
film on a metal [107]
Fig. 1.29 A surface chelate

formed between a catecholate
and the oxide film on a metal
in various environments. The inhibition of the pitting of aluminum in chloride

solutions was studied using substituted catechols [90]. A current interest in chelate
inhibitors has been prompted by the search to replace chromate inhibitors with
compounds which are less toxic and less polluting.
Hard and Soft Acids and Bases
Aramaki [7, 113–117] has given considerable attention to the study of corrosion
inhibition in terms of hard and soft acids and bases. Hard and soft acids and bases
(HSAB) are sub-sets of Lewis acids and bases. The designation “hard” refers to the
situation where the valence electrons of a molecule or ion are attracted by the
nucleus, and “soft” refers to the situation where the valence electrons are weakly
attracted to the nucleus. See Table 1.6 for properties of hard and soft acids and
bases and Table 1.7 for examples of common hard and soft acids and bases.
Aramaki [114, 117] explained passive film breakdown and repair by the HSAB
principle. Aramaki noted that Feþ3 is a hard acid (Table 1.7) so that passive films
containing Feþ3 ions are damaged by interaction with hard bases, such as Cl . The
halide ions Br (a base of intermediate hardness) and l (a soft base) would be less
Table 1.6 Classification of hard and soft acids and bases

Acid Base
Hard Electron cloud non-polarizable Electron cloud non-polarizable
Small size Small size
High positive charge
Soft Electron cloud more polarizable Electron cloud more polarizable
Large size Large size
Low positive charge
Table 1.7 Examples of hard and soft acids and bases [118]
Acids Bases
Hard H+, Li+, Na+, K+, Be+2, Mg+2, Ca+2, H2O, OH−, F−, CH3 CO2 , PO4 3 ,
Sr+2, Mn+2, Al+3, Ga+3, In+3, La+3, SO4 2 , Cl−, CO3 2 , ClO4 , NO3 ,
Cr+3, Co+3, Fe+3, Si+4, Ti+4, Zr+4, ROH, RO−, R2O, NH3, RH2
Th+4, U+4, AlCl3, Cr+6
Soft Cu+, Ag+, Au+, Tl+, Hg+, Pd+2, Cd+2, R2S, RSH, RS−, I−, SCN−, S2 O3 2 ,
Pt+2, Tl+3, GaCl3, I2, Br2, CN−, RNC, C2H4, C6H6
trinitrobenzene, quinones, CH2, Mo
(metal atoms)
Intermediate Fe+2, Co+2, Ni+2, Cu+2, Zn+2, Pb+2, C6H5NH2, C5H5N, Br−, NO2 ,
Sn+2, Sb+3, Bi+3, Rh+3, Ir+3, Ru+2, SO3 2
Os+2
Fig. 1.30 Schematic representation of the inhibition of pit initiation or the its propagation
according to the hard and soft acids and bases concept [113–117]
effective in causing film breakdown (although Br and I ions are larger than the
Cl ion and thus have more difficulty in penetrating the passive film).
Aramaki reasoned that effective corrosion inhibitors of passive film breakdown
should be Lewis bases whose hardness is greater that that of Cl . See Fig. 1.30. The
following hard bases were found to be effective in increasing the pitting potential in
a borate buffer containing 0.05 M Cl−:
OTP (octyl thiopropionic acid) C8 H17 SðCH2 Þ2 CO2

DS (N-dodecanoyl sarcosinate ion) C11 H23 CONðCH3 ÞCH2 CO2
Ol (oleic acid ion) C17 H33 CO2
Bz (benzoic acid) C6 H5 CO2
See Fig. 1.31 which shows a typical anodic polarization curve marking the ini-
tiation of pits at the pitting potential Epit and pit repassivation at the potential Erep.
To inhibit the propagation of pits, i.e., pit growth, Aramakii reasoned that soft
base inhibitors were required by the HSAB principle to interact with the underlying
metal, Feo , which is a soft base, according to Table 1.7. See Fig. 1.30. Aramaki and
coworkers [114, 117] investigated pit repassivation by measuring polarization
curves in simulated pit environments (high in Cl ion concentration and low in pH
Fig. 1.31 Schematic

representation of an anodic
polarization curve marking
the initiation of pits at the
pitting potential Epit and pit
repassivation at the potential
Erep. Eoc refers to the open
circuit (corrosion) potential.
Redrawn from Fujioka et al.
[117] by permission of
Elsevier
values) or by measuring the repassivation potential, shown schematically in

Fig. 1.31. Soft bases which were observed to be effective against pit propagation
were:
TU (thiourea) CSðNH2 Þ2
PTU (1-phenyl-2-thiourea) C6 H5 NHSCNH2
DPTU (1, 3-diphenyl-2-thiourea) ðC6 H5 NHÞ2 CS
DBTU (1, 3-dibutyl-2-thiourea) ðC4 H9 NHÞ2 CS
Br bromide ion
I iodide ion
To inhibit both pit initiation and pit propagation Aramaki and coworkers [114,
117] tried various combinations of hard and soft bases. Not all combinations were
successful, but the following combinations inhibited both pit initiation and pit
propagation:
OTP þ I
OTP þ DBTU
Bz þ DBTU (one combination)
Some of these effects are shown in Fig. 1.32. Other combinations raised Epit only
but not Erep , as is also shown in Fig. 1.31.
Fig. 1.32 Relationship between the pitting potential Epit and the repassivation potential Erep for
iron in Cl− containing borate buffer in the presence of various mixtures of hard and soft base
inhibitors. See the text for the key to abbreviations. Redrawn from Aramaki et al. [114, 117] by
Interaction of Cations with Oxide Films
Because Cl− anions interact with an ¼ AlOH2 þ surface at pH 7, then by the same
reasoning aqueous cations, e.g. Naþ , should interact with an ¼ AlO surface at
basic pH values. This interaction would produce a positively charged surface
consisting of ¼ AlO Naþ groups, which in turn would attract Cl anions, so that
pitting would easily occur at pH values above the isoelectric point. However, as
mentioned earlier, the pitting potential of aluminum increases above the isoelectric
point, so that a sequence of events involving an ¼ AlO Naþ Cl surface species
does not occur.
Instead, it is known that various cations are inhibitors for the corrosion of
aluminum and other metals [119–122]. In particular, cerium salts are effective
corrosion inhibitors for aluminum and its alloys.
Aramaki [115, 116] has pointed out that if an alkaline solution contains metallic
cations, the solubility product Ksp of the hydroxide (generally a small value) is
easily exceeded. Thus, a hydroxide is precipitated onto the existent oxide film and
this hydroxide interferes with the oxygen reduction reaction. In addition, the
hydroxide changes gradually to an oxide. For example, if a basic solution contains
dissolved Ceþ3 ions:
Ceþ3 þ 3 OH ! CeðOHÞ3 ðKsp ¼ 1:6 1020 Þ ð1:47Þ

Fig. 1.33 The solubility

product of various hydroxides
as a function of the hardness
index of the cation. Plotted
from data of Aramaki [113,
115]
and then:
CeðOHÞ3 ! CeO2 þ H2 O + Hþ þ e ð1:48Þ
Evidence that the prevailing oxidation state of precipitated cerium in alkaline

aerated solutions is +4 has been given by Aldykewicz et al. [123].
Aramaki [113, 115] has treated the precipitation reaction, Eq. (1.47), in terms of
Hard and Soft Acids and Bases (HSAB). For instance, inhibition of aluminum by
cerum salts [120, 121] can be explained by the HSAB principle. The cation Ceþ3 is
classified as a hard acid [7] and the anion OH as a hard base, so by the HSAB
principle, the precipitation reaction in Eq. (1.47) is favored. Aramaki [113, 115] has
shown that the logarithm of the solubility product for many hydroxides is a linear
function of the hardness index of the cation. See Fig. 1.33. Thus, the first step in the
formation of a precipitated oxide is also subject to Lewis acid-Lewis base
considerations.
Relationship Between the Isoelectric Point (pHpzc)

and the Potential of Zero Charge (Epzc) for Passive Metals
General Remarks
The surface charge on passive oxide films can be varied in two different ways: first
by changing the electrode potential, and second by changing the pH of the aqueous
solution, as shown schematically in Fig. 1.34.
Fig. 1.34 Schematic

illustration as to two different
approaches to change the
surface charge on a passive
oxide film
The isoelectric point of the oxide film (pHpzc) reflects how the charge on an
oxide film responds to changes in pH of the solution, whereas the potential of zero
charge (Epzc) determines how the charge on an oxide film responds to changes in
electrode potential of the oxide-covered metal. The relationship between these two
important parameters has been examined recently [124] and is considered below. So
far in this communication, we have considered isoelectric points but have not
considered potentials of zero charge for passive metals.
It should be noted that the general term “point of zero charge” is imprecise and
requires definition. This term is often used by colloid chemists to denote the pH
value of the aqueous solution in which a surface (usually an oxide) attains a net zero
charge, as described earlier (i.e., the isoelectric point).
Electrochemists, however, usually consider the “point of zero charge” to be the
potential of zero charge (Epzc), i.e., the electrode potential at which a metal/solution
interface displays a net zero charge. Thus, the term “potential of zero charge” by
convention refers to the bare (oxide-free) metal surface; and approaches such as
scraping, abrasion, laser ablation, or judicious choice of electrolyte are employed to
ensure that the metal surface is oxide-free. Indeed, the presence of an oxide film on
the metal surface, is sometimes considered to be a bane and the experiment is
thought to be compromised if the metal is oxide-covered. However, it is precisely
the condition of an oxide-covered metal that is of interest in the study of passivity.
The isoelectric point (pHpzc) and the potential of zero charge (Epzc) of oxide-
covered metals have the following properties:
1. The potential of zero charge (Epzc or Efb) is not a single quantity but varies with
the pH according to [125, 126]:
E Epzc ¼ 0:0591ðpH pHpzc Þ ð1:49Þ

Relationship Between the Isoelectric Point (pHpzc) … 41
2. The electrode potential at the isoelectric point of the oxide-covered surface

(pHpzc) is the potential of zero charge (Epzc) of the oxide-covered surface:
Epzc ¼ E at pH ¼ pHpzc ð1:50Þ
Measurement of Potentials of Zero Charge for Passive Metals
Potentials of zero charge of oxide-covered metals can be determined from their

semiconductor properties. A useful measure of the potential of zero charge of an
oxide-covered metal surface is the flatband potential, Efb . At the flatband potential
there is no band bending due to surface charges, so that this potential is also the
potential of zero charge. That is, Efb ¼ Epzc for oxide-covered metals.
The flatband potential can be measured from impedance analysis or from pho-
toelectrochemical experiments. In impedance measurements, the flatband potential
follows from Mott-Schottky plots [127]:
1 8p kT
2
¼ ½E Efb ð1:51Þ
Csc e e Nsc e
where CSC is the capacitance of the space charge region, NSC is the carrier density,
and the other symbols have their usual meanings, A plot of 1/C2 versus E gives a
straight line, which extrapolated to 1/C2 = 0 gives Efb þ kT=e, and hence Ffb.
A second method of determining Efb is from photoelectrochemical experiments,
in which no photocurrent is generated at the flatband potential [127]. Thus, the
photocurrent changes sign in passing through the flatband potential.
Potential Drop in the Metal/Oxide/Electrolyte System
The system under consideration is shown in Fig. 1.35. The electrode potential E
versus some reference electrode of an oxide-covered metal immersed in an elec-
trolyte is related to the inner potentials across the various phases or at interfaces.
That is [126, 128, 129]:
E ¼ D/SC þ D/H þ D/GC þ constant ð1:52Þ
where ϕSC refers to the space charge region of the oxide semiconductor, ϕΗ is the
inner potential in solution within the Helmholtz plane, and ϕGC is the inner potential
at the juncture of the outer Helmholtz plane and the beginning of the Gouy-Chapman
(diffuse region) of the electrical double layer. The term “constant” includes terms
which are not pH-dependent, such as the potential drop at the juncture of the metal
and its oxide film and the potential drop across the interior of the oxide film.
Fig. 1.35 Various interfaces present in the system: metal/oxide/solution. ϕSC refers to the innner
potential at the outermost space charge region, ϕH is the inner Helmholtz potential, ϕGC the inner
potential in the Gouy-Chapman diffuse double layer, and ϕsoln the inner potential in the bulk of the
aqueous solution. Adapted from Memming and Schwandt [129]
We are interested in the variation of each term in Eq. (1.52) with the pH of the
aqueous solution. Thus,
@E @D/SC @D/H @D/GC

¼ þ þ ð1:53Þ
@pH @pH @pH @pH
It has been shown previously [124] that the last term in Eq. (1.53) is small and
can be neglected. At the isoelectric point, pH ¼ pHpzc and the electrode potential
E is the potential of zero charge, Epzc. Thus, Eq. (1.53) becomes:
@Epzc @D/SC @D/H

¼ þ ð1:54Þ
@pHpzc @pHpzc @pHpzc
Thus, there are two sets of surface reactions which must be considered. The first
set of reactions involves the space charge side of the oxide/solution interface and
the formation and/or maintenance of the surface hydroxyl groups. The second set of
surface reactions involves the fate of the surface hydroxyl groups at the solution
side of the oxide/solution interface in terms of their association or dissociation.
Space Charge Side of the Oxide/Solution Interface
Both metallic cations Mþn (solid) and oxide anions O2 (solid) exist inside the
space charge layer of the oxide at the oxide/solution interface and can interact with
aqueous species (Hþ or OH ) to form or maintain the outermost layer of surface
hydroxyl ions, as mentioned earlier. Although, the space charge region extends
inward into the bulk of the oxide semiconductor, metallic cations or oxide anions in
the bulk of the space charge region cannot interact with aqueous Hþ or OH ions.
Thus, only that portion of the space charge region located near the oxide/solution
interface will contribute to the pH-dependence of /SC . That is,
@D/SC @ð/SC /b Þ @/SC

¼ ¼ ð1:55Þ
@pH @pH @pH
where /b is defined in Fig. 1.35.

Reactions of Hþ or OH ions with Mþn or O2 in the space charge region near
the oxide/solution interface can be written for either p-type or n-type oxides. We
first consider p-type oxides. For the interaction of Hþ ions with oxide anions:
O2 s þ Hþ aq þ hþ s ! OHsurf ð1:56Þ
where the subscripts s, aq, and surf refer to the solid, aqueous and surface phases,
respectively, and h+ refers to a hole. This reaction is shown schematically in
Fig. 1.36.
The condition for equilibrium in Eq. (1.56) is:
ðO2
l ðHþ
s Þþl ðhþ
aq Þ þ l ðOHsurf Þ
s Þ ¼ l ð1:57Þ
i is the electrochemical potential given by:

where l
i ¼ lO
l i þ RT ln ai þ zF/ ð1:58Þ
Fig. 1.36 Surface reaction on

a p-type oxide between a
hydrogen ion in solution and
an oxide site
Table 1.8 Types of interactions between an oxide semiconductor and solution species
Type oxide Interaction on Example
p-type Anion site O2 s þ Hþ aq þ hþ s ! OHsurf
p-type Cation site Niþ2 s þ OH aq þ hþ s ! Niþ2 s þ OHsurf
n-type Anion site O2 s þ Hþ aq ! OHsurf þ e s
n-type Cation site Znþ2 s þ OH aq ! Znþ2 s þ OHsurf þ e s
The subscripts “s”, and “aq” refer to solid and aqueous phases, and “surf” refers to the surface
where loi is the standard chemical potential, ai is the activity, and z is the charge for
charged surface species or ions located outside the Helmholtz plane (with z = 0 for
uncharged particles). The term / is the inner potential in the appropriate region.
Taking care to use the proper inner potential for each species, as in Fig. 1.36, leads
to:
constant þ 2:303 RT log½Hþ aq ¼ Fð/sc /soln Þ ð1:59Þ
or:

@D/sc 2:303RT
¼ ð1:60Þ
@pH F
where D/sc ¼ / sc /soln is the electrostatic potential in the space charge region
relative to the bulk solution.
There are three other cases to consider, and these are listed in Table 1.8. Each of
these three additional cases gives the same result as in Eq. (1.60) [124]. These
surface reactions are modifications of those proposed by Matsumoto et al. [130,
131].
Solution Side of the Oxide/Solution Interface
The behavior of surface hydroxyl groups has been discussed earlier in this com-
munication. The condition for equilibrium in Eq. (1.1) is:
l ðHþ
ðMOHsurf Þ þ l ðMOHþ
aq Þ ¼ l 2 surf Þ ð1:61Þ
and for Eq. (1.2)
l ðMO surf þ l
ðMOHsurf Þ ¼ l ðHþ
aq Þ ð1:62Þ
Use of Eq. (1.58) in Eq. (1.61) gives:
F(/H /soln Þ ¼ ½lo ðMOHsurf Þ þ lo ðHþ þ

aq Þ l ðMOH)2 surf
o
ð1:63Þ
þ RTfln½MOHsurf þ ln½Hþ þ
aq ln½MOH2 surf g
where concentrations (surface or bulk) are used in place of activities, /H is the inner
potential in the Helmholtz layer, and /soln is the inner potential in the bulk of the
solution. But /H /soln ¼ D/H , the potential drop across the Helmholtz layer
relative to the bulk of solution. In addition:
þ
lo ðMOHsurf þ lo ðHþ
aq Þ l ðMOHÞ2 surf Þ ¼ DG1 ¼ RT ln K1
o
ð1:64Þ
where K−1 is the acid dissociation constant for the reverse direction in Eq. (1.1).
Thus, Eq. (1.63) reduces to:
½MOHsurf ½Hþ
aq
FD/H ¼ RT ln K1 þ RT ln ð1:65Þ
½MOH2 þ surf
Similar considerations applied to Eq. (1.62) yield the result:
½MOH þ
surf ½Haq
FD/H ¼ RT ln K2 RT ln ð1:66Þ
½MOHsurf
where K2 is the acid dissociation constant for the forward direction in Eq. (1.2).
Simultaneous solution of Eqs. (1.65) and (1.66) by elimination of the term
[MOHsurf] gives [132, 133]:
2FD/H ½MOH2 þ surf

¼ logðK1 K2 Þ þ log þ 2 pH ð1:67Þ
2:303 RT ½MO surf
At the isoelectric point when rsurf ¼ 0; ½MOH2 þ surf ¼ [MO surf , the pH is the
isoelectric point pHpzc, and the quantity D/H at the isoelectric point can be denoted
by D/H . Thus, Eq. (1.67) becomes:
2:303 RT 2:303 RT
H ¼
D/O logðK1 K2 Þ pHpzc ð1:68Þ
2F F
and:

@D/H @D/O 2:303 RT
¼ H
¼ ð1:69Þ
@pH pzc @pHpzc F
Table 1.9 Values of the isoelectric point of oxides or oxide films and of the potential of zero
charge of oxide-covered metals which have n-type semiconductor oxides
Passive Isoelectric point Measured pH of Method of Epzc at pHpzc
metal (pHpzc) Epzc (V SCE) measurement measurement (V SCE)
of Epzc
Tungsten 2.5 [134] 0.30 [136] 1.3 M-S 0.21
(bulk oxide) 0.57 [137] 0.3 M-S 0.42
*3 [135]
0.25 [138] 1.0 M-S 0.14
(2.8 avg)
(0.26 avg)
Platinum 3.3 [44] 0.12 [139] 3.3 Capacitance 0.12
(passive film) minimum
Fe-20Cr 3–4.8 (3.9 avg) −0.44 [140] 9.2 M-S −0.13
Fe-12Cr [15] −0.32 [141] 8.4 M-S −0.05
(passive film)
(−0.09 avg)
Niobium 4.1 [53] (bulk oxide) −0.16 [142] 4 M-S −0.16
0.2 [143] 0.6 PE −0.01
(−0.09 avg)
316 4.4–5.2 −0.5 [144] 7.8 M-S −0.32
stainless (4.7 avg) [42] −0.43 [145] 8.35 M-S −0.21
steel (passive film)
(–0.27 avg)
Titanium 5.0 [31, 32] −0.24 [146] 0 M-S −0.54
(passive film) −0.23 [142] 5 M-S −0.23
−0.83 [147] 14 PE −0.30
−0.31 [148] 1.0 M-S −0.55
−0.10 [149] 0 M-S −0.40
−0.20 [150] 0 M-S −0.50
−1.13 [151] 14.7 M-S −0.56
−0.3 [152] 7 PE −0.18
0.53 [153] 7 PE −0.41
(−0.41 avg)
Tin 5–6 −0.25 [154] 8.5 M-S −0.073
(5.5 avg) [45] −0.19 [155] 6.0 M-S −0.16
(bulk oxide)
(−0.12 avg)
Zirconium 5.9 avg of values −1.10 [156] 5.9 PE −1.10
compiled by −1.25 [157] 8.5 PE −1.10
Natishan [41]
−1.13 [147] 14 PE −0.65
(bulk oxide)
−0.83 [147] 0.3 PE −0.83
−0.6 [158] 8.5 PE −0.45
−1.7 [159] 13 MS −1.20
(−0.96 avg)
Hafnium 7.1 [53] 0.0 [160] 0.0 PE −0.42
(bulk oxide)
Cobalt 7.8 [161] −0.64 [162] 8.3 Capacitance −0.58
(bulk oxide) minimum
(continued)
Table 1.9 (continued)

Passive Isoelectric point Measured pH of Method of Epzc at pHpzc
metal (pHpzc) Epzc (V SCE) measurement measurement (V SCE)
of Epzc
Aluminum 9.5 [17] −0.80 [163] 7 M-S −0.94
(passive film) −0.92 [164] 8.4 M-S −0.98
−0.81 [165] 7 PE −0.96
−1.40 [166] 7 M-S −1.54
(−1.11 avg)
Iron 9.9 [24, 30] −0.40 [167] 8.4 PE −0.49
(passive film) −0.63 [147] 14 PE −0.38
−0.37 [168] 9.9 M-S −0.37
−0.35 [169] 8.4 M-S −0.44
−0.5 [170] 9.2 M-S −0.54
−0.35 [171] 8.3 M-S −0.44
−0.32 [172] 8.4 M-S −0.41
−0.46 [173] 9.4 M-S −0.49
−0.44 [174] 9.2 M-S −0.48
(−0.45 avg)
Zinc 9.3 [23] −0.6 [176] 10 M-S −0.54
(passive film) −1.13 [177] 5 M-S −1.37
8.7 [175]
−0.61 [178] 8.9 M-S −0.62
(bulk oxide)
(9.0 avg) −1.13 [147] 13 PE −0.73
−0.80 [179] −0.8 PE −0.92
(−0.84 avg)
Lead 11.3 [180] −1.27 [181] 13 M-S −1.17
(bulk oxide) −0.58 [182, 0 M-S −1.25
183]
(−1.21 avg)
*
M-S Mott-Schottky, PE photoelectrochemistry
Thus, use of Eqs. (1.69) and (1.60) in:
@Epzc @D/SC @D/H

¼ þ ð1:54Þ
@pHpzc @pHpzc @pHpzc
gives:
@Epzc 2:303 RT 2:303 RT

¼ ¼ 0:120 V ð1:70Þ
@pHpzc F F
Table 1.10 Values of the isoelectric point of oxides or oxide films and of the potential of zero
charge of oxide-covered metals which have p-type semiconductor oxides
Passive Isoelectric Measured pH of Method of Epzc at
metal point (pHpzc) Epzc (V SCE) measurement measurement pHpzc
of Epzc (V SCE)
Chromium 5.3 [17] 0.50 [182, 183] 5.3 M-S 0.50
(passive film) 0.67 [182, 183] 5.3 M-S 0.67
0.35 [184] 8.4 M-S 0.53
0.5 [185] 9.2 M-S 0.74
0.70 [186] 8.5 M-S 0.89
0.46 [187] 9.2 M-S 0.70
0.39 [174] 9.2 M-S 0.62
(0.66 avg)
Nickel 8.4 [188] 0.3 [189] 8.4 PE 0.3
(bulk oxide) 0.4 [190] 8.4 M-S 0.4
−0.14 [191] 14.5 M-S 0.22
0.12 [192] 14.8 M-S 0.49
0.72 [174] 9.2 M-S 0.76
0.10 [148] 14 M-S 0.43
(0.43 avg)
Copper 9.1 [23] −0.02 [147] 14 PE 0.24
9.8 [193] 0.01 [194] 13 PE 0.22
(9.5 avg)
0.16 [195] 9.8 PE 0.12
(passive film)
−0.275 [196] 12.85 PE −0.07
(0.13 avg)
*
M-S Mott-Schottky; PE photoelectrochemistry
Numerical Evaluation
Tables 1.9 and 1.10 list literature values for the isoelectric point pHpzc for oxide
films on various passive metals or bulk oxides. Tables 1.9 and 1.10 also list values
for the flatband potential (Efb = Epzc) for various semiconductor oxide films. These
values were taken mostly from Mott-Schottky plots or from photoelectrochemical
measurements. The value of Epzc at the pH of measurement was converted to the pH
of the isoelectric point, pHpzc, by Eq. (1.49).
Figure 1.37 is a plot of Epzc at the isoelectric point versus the value of the
isoelectric point, pHpzc for various metals having passive films which are oxide
semiconductors. It can be seen that there is considerable scatter in the data, but that
the relationship is linear, with dEpzc/dpHpzc = −0.142 V for n-type oxides and
−0.115 V for p-type oxides. Both values are in nominal agreement with the value
−0.120 V, as predicted by Eq. (1.70).
Reasons for the scatter in Fig. 1.37 are due to difficulties in either measurement
or modeling and have been discussed in detail elsewhere [124]. The correlation
Fig. 1.37 Relationship between the isoelectric point (pHpzc) and the potential of zero charge (Epzc
or Efb) for various passive metals. The open circles are the average values of Efb for n-type oxide
films from Table 1.9. The closed circles are average values of Efb for p-type oxide films from
Table 1.10. The range of values for Efb is also shown for each passive metal. Reproduced from
Ref. [124] by permission of Elsevier
coefficient for the linear fit is fairly large (r = 0.84 for n-type oxides and r = 0.94 for
p-type oxides. The observed slopes of Epzc versus pHpzc in Fig. 1.37 of −0.142 V
for n-type oxides and −0.115 V for p-type oxides are in nominal agreement with the
theoretical value −0.120 V.
Summary
The acid-base properties of oxide films on metal surfaces determine the surface
charge character of the passive film. The surface charge character, in turn, is
important in the surface phenomena of pitting, adhesion of polymers, corrosion
inhibition, and the adsorption of inhibitive or aggressive ions from solution. Several
illustrative examples have been given. The relationship between the isoelectric
point and the potential of zero charge for passive films has also been considered.
References
1. E. McCafferty, A.C. Zettlemoyer, Disc. Faraday Soc. 52, 239 (1971)

2. W. Stumm, Chemistry of the Solid-Water Interface (Wiley, New York, 1992)
3. E. McCafferty, J. Electrochem. Soc. 150, B342 (2003)
4. E. McCafferty, J.P. Wightman, Surf. Interface Anal. 26, 549 (1998)
5. W.F. Luder, Chem. Revs. 27, 547 (1940)
6. G.N. Lewis, J. Franklin Inst. 226, 293 (1938)

7. K. Aramaki, Corrosion Engineering 45, 733 (1996)
8. T. Yamanaka, K. Tanabe, J. Phys. Chem. 80, 1723 (1976)
9. M. Kaltchev, W.T. Tysoe, Surf. Sci. 430, 29 (1999)
10. H. Ma, Y. Berthier, P. Marcus, Corros. Sci. 44, 171 (2002)
11. D. Fairhurst, V. Ribitsch, Particle Size Distribution II. in ed. by T. Provder, ACS Symposium
Series No, 472, p. 337, (American Chemical Society, Washington, DC, 1991)
12. C. Bellmann, A. Opfermann, H.-J. Jacobasch, H.-J. Adler, Fresnius. J. Anal. Chem. 358, 255
(1997)
13. S. Roessler, R. Zimmerman, D. Scharnweber, C. Werner, H. Worch, Colloids Surf. B,
Biointerfaces 26, 387 (2002)
14. B.R. Strohmeier, Surf. Interface Anal. 15, 51 (1990)
15. L. Boulangé-Petermann, A. Doren, B. Baroux, M.-N. Bellon-Fontaine, J. Colloid Interface
Sci. 171, 179 (1995)
16. L.-K. Chau, M.D. Porter, J. Colloid Interface Sci. 145, 283 (1995)
17. E. McCafferty, J.P. Wightman, J. Colloid Interface Sci. 194, 344 (1997)
18. E. McCafferty, J.P. Wightman, J. Adhe. Sci. Technol. 13, 1415 (1999)
19. E. McCafferty, J. Electrochem. Soc. 146, 2863 (1999)
20. W.A. Adamson, Physical Chemistry of Surfaces, 5th edn. (Wiley, New York, 1990). 386
21. P. Atkins, Physical chemistry (W. H. Freeman and Co., New York, NY, 1998). 161
22. S. Scotto-Sheriff, E. Darque-Ceretti, G. Plassart, M. Aucouturier, J. Mater. Sci. 34, 5081
(1999)
23. D. Chvedov, E.L.B. Logan, Colloids Surf. A 240, 211 (2004)
24. G.W. Simmons, B.C. Beard, J. Phys. Chem. 91, 1143 (1987)
25. J.F. Watts, E.M. Gibson, Int. J. Adhes. Adhe. 11, 105 (1991)
26. G.A. Parks, Chem. Revs. 65, 177 (1965)
27. R.D. Shannon, C.T. Prewitt, Acta Cryst. B25, 925 (1969)
28. M. Delamar, J. Electron. Spectr. Rel. Phenom. 53, c11 (1990)
29. C. Sun, J.C. Berg, Adv. Colloid Interface Sci. 105, 151 (2003)
30. G. Kurbatov, E. Darque-Ceretti, M. Aucouturier, Surf. Interface Anal. 18, 811 (1992)
31. E. McCafferty, J.P. Wightman, Appl. Surf. Sci. 143, 92 (1999)
32. E. McCafferty, J.P. Wightman, T. Frank, Cromer. J. Electrochem. Soc. 146, 2849 (1999)
33. K. Hu, F-R. F. Fan, A.J. Bard, A.C. Hillier J. Phys. Chem. B. 101, 8298 (1997)
34. I. Larson, C.J. Drummond, D.Y.C. Chan, F. Grieser, J. Am. Chem. Soc. 115, 11885 (1993)
35. M.G. Cattania, S. Ardizzone, C.L. Bianchi, S. Carella, Colloids Surf. A 76, 233 (1993)
36. C.E. McNamee, Y. Tsujii, M. Matsumoto, Langmuir 21, 11283 (2005)
37. P.G. Hartley, I. Larson, P.J. Scales, Langmuir 13, 2207 (1997)
38. S. Veeramasuneni, M.R. Yalamanchili, J.D. Miller, J. Colloid Interface Sci. 184, 594 (1996)
39. M.S. Lim, S.S. Perry, H.C. Galloway, D.C. Koeck, J. Vac. Sci. Technol. B 20, 575 (2002)
40. S. Ardizzone, S. Trasatti, Adv. Colloid Interface Sci. 64, 173 (1996)
41. P.M. Natishan, E. McCafferty, G.K. Hubler, J. Electrochem. Soc. 135, 321 (1988)
42. N. Kallay, D. Kovacevic, I. Dedic, V. Tomasic, Corrosion 50, 598 (1994)
43. H. Tamura, T. Oda, M. Nagayama, R. Furuichi, J. Electrochem. Soc. 136, 2782 (1989)
44. N. Kallay, Z. Torbic, M. Golic, E. Matijevic, J. Phys. Chem. 95, 7028 (1991)
45. T. Arai, D. Aoki, Y. Okaba, M. Fujihara, Thin Solid Films 273, 322 (1996)
46. W.P. Hsu, L. Rönnquist, E. Matijevic, Langmuir 4, 31 (1988)
47. M. Ozawa, M. Hattori, J. Alloys Compd. 408–412, 560 (2006)
48. M. Mao, D. Fornasiero, J. Ralston, R. St. C. Smart, S. Sobieraj, Colloids Surf. A, 85, 37
(1994)
49. H. Moriwaki, Y. Yoshikawa, T. Morimoto, Langmuir 6, 847 (1990)
50. G.A. Kokarev, V.A. Kolesnikov, A.F. Gubin, A.A. Korabonav, Soviet Electrochem. 18, 407
(1982)
51. D.E. Yates, T.W. Healy, J. Colloid Interface Sci. 52, 222 (1975)
52. G.W. Smith, T. Salman, Canadian Met. Quart. 5(2), 93 (1966)
References 51
53. M. Kosmulski, Langmuir 13, 1615 (1997)

54. W. Janusz, J. Colloid Interface Sci. 145, 119 (1991)
55. J.R. Galvele, in Passivity of Metals. ed by R.P. Frankenthal, J. Kruger (The Electrochemical
Society Princeton, NJ 1978), p. 285
56. J.H. de Boer, H. Verwey in Advances in Colloid Science, ed. by Mark, E.J.W. (Interscience
Publishers, New York, 1950) Vol. III, Chapter 1, p. 1
57. A. Kolics, A.S. Besing, P. Baradlai, R. Haasch, A. Weickowski, J. Electrochem. Soc. 148,
B251 (2001)
58. M. Kosmulski, Colloids Surf. A 83, 237 (1994)
59. E. McCafferty, Corros. Sci. 45, 1421 (2003)
61. L. Tomcsanyi, K. Varga, I. Bartik, G. Horanyi, E. Maleczki, Electrochim. Acta 34, 855
(1989)
62. G. Okamato, T. Shibata, in Proceedings of the Third International Congress on Metallic
Corrosion. Vol. 1, p. 396, Moscow (1969)
63. O.J. Murphy, J.O’M. Bockris, T.E. Pou, D.L. Cocke, G. Sparrow, J. Electrochem. Soc. 129,
2149 (1982)
64. W.M. Carroll, C.B. Breslin, Br. Corros. J. 26, 255 (1991)
65. P.M. Natishan, W.E. O’Grady, E. McCafferty, D.E. Ramaker, K. Pandya, A. Russell,
J. Electrochem. Soc. 146, 1737 (1999)
66. F.D. Bogar, R.T. Foley, J. Electrochem. Soc. 119, 462 (1972)
67. S. Dallek, R.T. Foley, J. Electrochem. Soc. 123, 1775 (1976)
68. J.R. Galvele, S.M. de De, Michele. Corros. Sci. 10, 795 (1970)
69. B.N. Stirrup, N.A. Hampson, I.S. Midgley, J. Appl. Electrochem. 5, 229 (1975)
70. H. Bohni, H.H. Uhlig, J. Electrochem. Soc. 116, 906 (1969)
71. H. Kaesche, Z. Physik Chem. N. F. 34, 87 (1962)
72. Z.A. Foroulis, M.J. Thubrikar, Werkstoffe u. Korros. 26, 350 (1975)
73. J. O’M. Bockris and Lj. V. Minevski, J. Electroanal. Chem. 349, 375 (1993)
74. J. Schultz, M. Nardin in Adhesion Promotion Techniques. ed. by K.L. Mittal, A. Pizzi
(Marcel Dekker, New York 1999), p. 1
75. F.M. Fowkes, D.O. Tischler, J.A. Wolfe, L.A. Lannigan, C.M. Ademu-John, M.J. Halliwell,
J. Polymer Sci. Polymer Chem. Edition, 22, 547 (1984)
76. F.M. Fowkes, M.A. Mostafa, Ind. Eng. Chem. Prod. Res. Dev. 17, 3 (1978)
77. F.M. Fowkes in Encyclopedia of polymer science and engineering. Supplement Volume
(Wiley Interscience, New York 1989), p. 1
78. H.-J. Jacobasch, K. Grundke, S. Schneider, F. Simon, J. Adhes. 48, 57 (1995)
79. M.A. Tshabalala, J. Appl. Polym. Sci. 65, 1013 (1997)
80. C.J. van Oss, M. Chaudhury, R.J. Good, Chem. Rev. 88, 927 (1988)
81. C.J. van Oss, R.J. Good, J. Macromol. Sci.-Chem. A, 26, 1183 (1989)
82. R.J. Good, C.J. van Oss in Modern approaches to wettability. ed. by Schrader, M.E. Loeb, G.
I (Plenum Press, New York 1992), p. 1
83. R.J. Good, J. Adhes. Sci. Technol. 6, 1269 (1992)
84. A.D. Roberts, Langmuir 8, 1479 (1992)
85. J.W. Chin, J.P. Wightman, J. Adhes. 41, 23 (1993)
86. F. Bodino, N. Compiègne, L. Köhler, J.J. Pireaux, R. Caudano in Proceedings of the 20th
Annual Meeting of the Adhesion Society. (ed. by L.T. Drzal, H.P. Schreiber) p. 41 (1997)
87. Y. Liu, S.H. Goh, S.Y. Lee, C.H.A. Huan, J. Polymer Sci. B Polymer Phys. 38, 501 (2000)
88. J.F. Watts, M.M. Chehimi, E.M. Gibson, J. Adhesion 39, 145 (1992)
89. S.R. Leadly, J.F. Watts, J. Electron Spectr. Related Phenom. 85, 107 (1997)
90. D.C. Hansen, E. McCafferty, J. Electrochem. Soc. 143, 114 (1996)
91. E. McCafferty, J. Adhes. Sci. Technol. 16, 239 (2002)
92. I.A. Starostina, R.R. Khasbiullin, O.V. Stoyanov, A.E. Chalykh, Russian J. Appl. Chem. 74,
1920 (2001)
93. W. Funke, Prog. Org. Coat. 31, 5 (1997)
94. J.H. W. de Wit in Corrosion mechanisms in theory and practice. ed. by P. Marcus, J. Oudar
(Marcel Dekker, New York 1995), p. 581
95. H. Leidheiser, Jr. W. Funke, J. Oil Col. Chem. Assoc. 70(5), 121 (1987)
96. Y.-F. Yu, J.J. Chessick, A.C. Zettlemoyer, J. Phys. Chem. 63, 1626 (1959)
97. P.F. Rossi, M. Bassoli, G. Olivera, F. Guzzo, J. Thermal Anal. 41, 1227 (1994)
98. Y.-F. Yu Yao, J. Phys. Chem. 67, 2055 (1963)
99. M. Nakazawa, G.A. Somorjai, Appl. Surf. Sci. 84, 309 (1995)
100. Y.-F. Yu Yao, J. Phys. Chem. 69, 3930 (1965)
101. T. Matsuda, T. Taguchi, M. Nagao, J. Thermal Anal. 38, 183 (1992)
102. J.G.N. Thomas, Br. Corros. J. 5, 41 (1970)
103. J.E.O. Mayne, C.L. Page, Br. Corros. J. 7, 115 (1972)
104. B. Müller, W. Kläger, Corros. Sci. 38, 1869 (1996)
105. B. Müller, React. Funct. Polym. 39, 165 (1999)
106. F.P. Dwyer, D.D. Mellor, Chelating agents and metal chelates (Academic Press, New York,
NY, 1964)
107. A. Weisstuch, D.A. Carter, C.C. Nathan, Mater. Prot. Perf. 10(4), 11 (1971)
108. D.C. Zecher, Mater. Perf. 15(4), 33 (1976)
109. R.L. LeRoy, Corrosion 34, 98 (1978)
110. B. Müller, W. Kläger, G. Kubitzki, Corros. Sci. 39, 1481 (1997)
111. F. Tirbonod, C. Fiaud, Corros. Sci. 18, 139 (1978)
112. Y.I. Kuznetsov, T.I. Bardasheva, Russ. J. Appl. Chem. 66, 905 (1993)
113. K. Aramaki, Corr. Eng. 46, 389 (1997)
114. K. Aramaki, Corr. Eng. 46, 913 (1997)
115. K. Aramaki, Corrosion 55, 157 (1999)
116. K. Aramaki, Corros. Sci. 43, 1573 (2001)
117. E. Fujioka, H. Nishihara, K. Aramaki, Corros. Sci. 38, 1669 (1996)
118. R.G. Pearson, J. Chem. Ed. 45, 581 (1968)
119. W.R. Buck III, H. Leidheiser Jr, Corrosion 14, 308t (1958)
120. B.R.W. Hinton, D.R. Arnott, N.E. Ryan, Met. Forum 7(4), 211 (1984)
121. D.R. Arnott, B.R.W. Hinton, N.E. Ryan, Mater. Prot. 26(8), 42 (1987)
122. M.G.A. Khedr, A.M.S. Lashien, Corros. Sci. 33, 137 (1992)
123. A.J. Aldykewicz, A.J. Davenport, H.S. Isaacs, J. Electrochem. Soc. 143, 147 (1996)
124. E. McCafferty, Electrochim. Acta 55, 1630 (2010)
125. H. Gerischer, Electrochim. Acta 34, 1005 (1989)
126. S.R. Morrison, in Electrochemistry at semiconductor and oxidized metal electrodes. (Plenum
Press, New York 1980), p. 56–62
127. J.O’M. Bockris, S.U.M. Khan, Surf. Electrochem. (Plenum Press, New York, 1993),
pp. 179–182
128. J.O’M. Bockris, A.K.N. Reddy, Modern electrochemistry. Vol. 2 (Plenum Press, New York
1977), p. 673
129. R. Memming, G. Schwandt, Angew. Chem. Internat. Ed. 6, 851 (1967)
130. Y. Matsumoto, T. Yoshikawa, E. Sato, J. Electrochem. Soc. 136, 1389 (1989)
131. Y. Matsumoto, K. Sugiyama, E. Sato, J. Electrochem. Soc. 135, 98 (1988)
132. S. Levine, A.L. Smith, Disc. Faraday Soc. 52, 290 (1971)
133. D.E. Yates, S. Levine, T.W. Healey, J. Chem. Soc. Faraday Trans. I 70, 1807 (1974)
134. M. Anik, T. Consizoglu, J. Appl. Electrochem. 36, 603 (2006)
135. M. Anik, K. Osseo-Asare, J. Electrochem. Soc. 149, B224 (2002)
136. S.R. Biaggio, R.C. Rocha-Filho, J.R. Vilche, F.E. Varela, L.M. Gassa, Electrochim. Acta 42,
1751 (1997)
137. F. Di Quarto, A. Di Paola, C. Sunseri, Electrochim. Acta 26, 1177 (1981)
138. G. Vazquez, I. Gonzalez, J. Electrochem. Soc. 154, C702 (2007)
139. E. Gileadi, S.D. Argade, J.O’M. Bockris. J. Phys. Chem. 70, 2044 (1966)
140. J.-B. Lee, S.-W. Kim, Materials. Chem. Phys. 104, 98 (2007)
141. E.B. Castro, J.R. Vilche, Electrochim. Acta 38(1567), 775 (1993)
References 53
142. K.E. Heusler, K.S. Yun, Electrochim. Acta 22, 977 (1977)
143. M. Arita, Y. Hayashi, Mat Trans JIM 35, 233 (1994)
144. M. Da Cunha Belo, B. Rondot, C. Compere, M.F. Montemor, A.M. Simoes, M.G.S. Ferreira,
Corros. Sci. 40, 491 (1998)
145. I. Nicic, D.D. Macdonald, J. Nuclear Mater. 379, 54 (2008)
146. J.W. Schultze, U, Stimming, J. Weise, Ber. Bunsenges. Phys. Chem. 86, 276 (1982)
147. T.D. Burleigh, Corrosion 45, 464 (1989)
148. S.P. Harrington, T.M. Devine, J. Electrochem. Soc. 155, C381 (2008)
149. A.S. Bondarenko, G.A. Ragoisha, J. Solid State Electrochem. 9, 845 (2005)
150. D.-S. Kong and J.-X, Wu, J. Electrochem. Soc. 155, C32 (2008)
151. S.U.M. Khan, J. Akikusa, J. Electrochem. Soc. 145, 89 (1998)
152. T. Bachmann, W. Vonau, P. John, Anal. Bioanal. Chem. 374, 715 (2002)
153. P. Clechet, J.-R. Martin, R. Olier, C. Vallouy, C. R. Acad. Sci. Ser. C 887 (1976)
154. S. Kapusta, N. Hackerman, Electrochim. Acta 25, 949 (1980)
155. M. Metikos-Hukovic, S. Omanovic, A. Jukic, Electrochim. Acta 45, 977 (1999)
156. F. DiQuarto, S. Piazza, C. Sunseri, M. Yang, S.-M. Cai, Electrochim. Acta 41, 2511 (1996)
157. M. Santamaria, F. Di Quarto, H. Habazaki, Electrochim. Acta 53, 2272 (2008)
158. S.J. Lee, E.A. Cho, S.J. Ahn, H.S. Kwon, Electrochim. Acta 46, 2605 (2001)
159. J.-Y. Huot, J. Appl. Electrochem. 22, 852 (1992)
160. F. DiQuarto, M. Santamaria, P. Skeldon, G.E. Thompson, Electrochim. Acta 48, 1143 (2003)
161. S. Ardizzone, G. Spinolo, S. Trasatti, Electrochim. Acta 40, 2683 (1995)
162. D. Gallant, M. Pézolet, S. Simard, Electrochim. Acta 52, 4927 (2007)
163. L. Kobotiatis, N. Kioupis, P.G. Koutsoukos, Corrosion 53, 562 (1997)
164. J.O’M. Bockris, Y. Kang, J. Solid State Electrochem. 1, 17 (1997)
165. J.C.S. Fernandes, R. Picciochi, M. Da Cunha Belo, T. Moura e Silva, M.G.S. Ferreira, I.T.E.
Fonseca, Electrochim. Acta, 49, 4701 (2004)
166. F.J. Martin, G.T. Cheek, W.E. O’Grady, P.M. Natishan, Corros. Sci. 47, 3187 (2005)
167. P. Schmuki, H. Böhni, Electrochim. Acta 40, 775 (1995)
168. S.M. Wilhelm, N. Hackerman, J. Electrochem. Soc. 128, 1668 (1981)
169. E.B. Castro, J.R. Vilche, Electrochim. Acta 38(1567), 775 (1993)
170. L. Hamadou, A. Kadri, N. Benbrahim, Appl. Surf. Sci. 252, 1510 (2005)
171. Y.M. Zeng, J.L. Luo, Electrochim. Acta 48, 3551 (2003)
172. F.M. Delnick, N. Hackerman, J. Electrochem. Soc. 126, 732 (1979)
173. V. Horvat-Radosevic, K. Kvastek, Electrochim. Acta 42, 1403 (1997)
174. M. Da Cunha Belo, N.E. Hakiki, M.G.S. Ferreira, Electrochim. Acta. 44, 2473 (1999)
175. M. Valtiner, S. Borodin, G. Grundmeier, Langmuir 24, 5350 (2008)
176. A.E. Bohe, J.R. Vilche K. Jüttner, W.J. Lorenz, W. Kautek, W. Paatsch, Corros. Sci. 32, 621
(1991)
177. L. Sziraki, A. Cziraki, I. Gerocs, Z. Vertesy, L. Kiss, Electrochim. Acta 43, 175 (1998)
178. H. Mishima, B.A. Lopez de Mishima, E. Santos, C.P. De Pauli, K. Azumi, N. Sato,
Electrochim. Acta 36, 1491 (1991)
179. E.A. Thompson, T.D. Burleigh, Corros. Eng. Sci. Technol. 42, 237 (2007)
180. J. Biscan, M. Kosec, N. Kallay, Colloids Surf. A 79, 217 (1993)
181. S.-J. Xia, W.-F. Zhou, Electrochim. Acta 40, 175 (1995)
182. Z.-L. He, C. Pu, W.-F. Zhou, J. Power Sources 39, 225 (1992)
183. D.-S. Kong, S.-H. Chen, C. Wang, W. Yang, Corros. Sci. 45, 747 (2003)
184. S. Virtanen, P. Schmuki, H, Böhni, P Vuoristo, T. Mäntylä, J. Electrochem. Soc. 142, 3067
(1995)
185. M.G.S. Ferreira, M. Da Cunha Belo, N.E. Hakiki, G. Goodlet, M.F. Montemor, A.M.
P. Simoes, J. Braz. Chem. Soc. 13, 433 (2002)
186. J.S. Kim, E.A. Cho, H.S. Kwon, Electrochim. Acta 47, 415 (2001)
187. G. Goodlet, S. Faty, S. Cardosa, P.P. Freitas, A.M.P. Simoes, M.G.S. Ferreira, M. Da Cunha
Belo. Corros. Sci. 46, 1479 (2004)
188. G.A. Kokarev, V.A. Kolesnikov, A.F. Gubin, A.A. Korabonav, Soviet Electrochem. 18, 407
(1982)
189. H.J. Jang, C.J. Park, H.S. Kwon, Electrochim. Acta 50, 3503 (2005)
190. E. Sikora, D.D. Macdonald, Electrochim. Acta 48, 69 (2002)
191. M.J. Madou, M.C.H. McKubre, J. Electrochem. Soc. 130, 1056 (1983)
192. J. Kang, Y. Yang, X. Jiang, H. Shao, Corros. Sci. 50, 3576 (2008)
193. N. Kallay, Z. Torbic, E. Barouch, J. Jednacak-Biscan, J. Colloid Interface Sci. 118, 431
(1987)
194. U. Collisi, H.-H. Strehblow, J. Electroanal. Chem. 284, 385 (1990)
195. K. Nakaoka, J. Ueyama, K. Ogura, J. Electrochem. Soc. 151, C661 (2004)
196. F. Caballero-Briones J.M. Artes, I. Diaz-Perez, P. Gorostiza, F. Sanz, J. Phys. Chem. C 113,
1028 (2009)
Chapter 2
Capillarity and Corrosion
Introduction
The surface of a liquid or solid exhibits the property of surface tension because
molecules at the surface of the liquid or solid have fewer neighbors than those in the
bulk. This situation leads to a net force perpendicular to the surface plane toward
the interior of the liquid (or solid). In the case of solids, the property is called its
surface free energy and may depend on the orientation of the solid surface.
The surface tension of water at 20 °C is 72.8 mN/m (millinewtons/m). The
following units are equivalent:
72:8 dynes=cm ¼ 72:8 erg=cm2 ¼ 72:8 mN=m ¼ 72:8 mJ=m2
Some values of surface tension for various liquids are given in Table 2.1 [1, 2].
Capillarity
When a small-diameter glass tube is immersed in a liquid, the liquid will rise in the
capillary tubing, as shown in Fig. 2.1. (If the liquid is mercury there will be a
depression rather than a rise in the capillary column, but we will be dealing with
only aqueous fluids here.) The height of the capillary rise (h) is given by:
2c
h¼ ð2:1Þ
qgr

56 2 Capillarity and Corrosion
Table 2.1 Some surface

Water 72.94
tension values in mN/m at
20 °C [2] Methylene iodide 67.00
Dimethyl sulfoxide 43.54
Benzene 28.88
Methanol 22.50
Perfluoroheptane 13.19
Fig. 2.1 Capillary rise of

water in a small-diameter
glass tube
where γ is the surface tension of the liquid, ρ is the density of the wetting liquid, g is
the acceleration due to gravity, and r is the radius of the capillary.
This equation arises as follows. The force upward in the capillary liquid is:
FðupÞ ¼ cð2prÞ cos h ð2:2Þ
where θ is the contact angle and is usually zero (see Fig. 2.1).
The downward force is due to the weight of the liquid and is given by:
FðdownÞ ¼ pr2 qgh ð2:3Þ
At equilibrium, the force up and the force down are equal:
FðupÞ ¼ FðdownÞ ð2:4Þ
This gives Eq. (2.1).

Figure 2.2 shows some calculated values of the capillary rise h for various
capillary radii r. The smaller the capillary radius, the higher the capillary rise.
Application to Organic Coatings 57
Fig. 2.2 Calculated capillary

rise h (in mm) for various
glass tubing radii r (in mm)
Application to Organic Coatings
It is well-known that organic coatings do not form perfect barriers to corrosion, but
instead they are semi-permeable membranes into which water, oxygen, and chloride
ions can pass. Nguyen et al. [3] have detected the presence of molecular water in
organic coatings on silicon using Fourier transform infrared-multiple internal
reflection. Their results showing the accumulation of water at the organic coating/
metal interface are given in Fig. 2.3. Consider an organic coating which is 500 μm
Fig. 2.3 Thickness and amount of water at the coating/substrate interface for clear epoxy coatings
on silicon and silane-treated silicon. The epoxy coating thickness was 130–140 mm. Redrawn
from Nguyen et al. [3] with the permission of Elsevier
Fig. 2.4 a Schematic

drawing of an irregularly
shaped capillary path in an
organic coating. b Capillary
rise in a porous organic
coating
thick (0.5 mm) and having an array of pores which align to give an irregularly
shaped capillary path through the organic coating, as shown in Fig. 2.4.
If the accumulated liquid water at the base of the organic coating encounters the
porous path in the coating, it will be possible for liquid water at the base of the
coating to undergo capillary rise. For an effective pore radius of 3 μm, the height of
capillary rise is calculated from Eq. (2.1) to be h = 5.0 m. This means that the
capillary height far exceeds the thickness of the organic coating. See Table 2.2.
If the effective pore radius is 30 μm, the calculated height of capillary rise is
0.50 m. We can then consider that some of the organic matter from the coating can
leach out into solution, so that the surface tension of the liquid phase will decrease,
say to half its value for pure water, 36 mN/m. In that case, the calculated height of
capillary rise is still substantial at 0.24 m and exceeds the thickness of the organic
coating. Thus, the aqueous solution in the pores connects the internal electrolyte
within the network of pores to the external electrolyte, which may be a thin film of
water condensed from the atmosphere. This serves to promote degradation of the
organic coating and underfilm corrosion.
Table 2.2 Calculated capillary rise in an organic coating

Radius r (μm) Surface tension γ (mN/m) Calculated capillary rise h (m)
3 72.8 5.0
30 72.8 0.50
30 36 0.24
Young and Laplace Equation 59
Young and Laplace Equation
The pressure difference Dp across a spherical bubble of a fluid is given by the

Young and Laplace equation:
Dp ¼ 2c=r ð2:5Þ
where r is the radius of the spherical bubble. This equation arises by considering the
free energy change involved when a spherical bubble expands from radius r to
radius r + dr. The free energy change is:
DG ¼ c dA Dp dV ð2:6Þ
where A is the area of the bubble and V is its volume.

The minus sign in Eq. (2.6) is required because work is done by the system.
With A = 4πr2, dA = 8πrdr. With V = (4/3)πr3, dV = 4πr2dr. Substitution of these
simple expressions for dA and dV into Eq. (2.6) with DG = 0 at equilibrium gives
the result in Eq. (2.5).
If the curved surface is not spherical but instead has two principal radii of
curvature r1 and r2 as in Fig. 2.5, then the term 1/r in Eq. (2.5) is replaced by the
term (1/r1 + 1/r2), and the Young and Laplace equation becomes:
Fig. 2.5 A curved surface

defined by two different radii
of curvature. Adapted from
Ref. [1], p. 9
Fig. 2.6 The force required to separate two flat glass plates (separated by a thin film of liquid) can
be calculated from the Young and Laplace equation

1 1
Dp ¼ c þ ð2:7Þ
r1 r2
The derivation of this equation has been given by Adam [1] and Adamson [2].
When r1 = r2 = r, Eq. (2.7) reduces to Eq. (2.5).
Force Between Glass Plates
The Young and Laplace equation can be used to calculate the pressure required to
separate two flat glass plates which are held together by a thin film of water (or of
an aqueous solution). See Fig. 2.6. Here r1 is given by the radius of curvature of the
meniscus and r2 = infinity. Thus,

1
Dp ¼ c ð2:8Þ
r1
if the two glass plates are separated by a 0.5 mm thin layer of water, then Dp is
calculated to be Dp = 146 Pa. Note that the calculated pressure difference depends
only on the surface tension of the wetting liquid and on the geometry of the system.
The material properties of the solids are not involved in the calculation. The force
between the two glass plates is then: Force = Dp × (Wetted area of plates).
Application to Poultice Corrosion
Poultice corrosion is the name given collectively to corrosion that results from the
accumulation of dirt, dust, corrosion products, and other forms of debris.
When this happens, a highly corrosive condition can be created. Strictly
speaking, poultice corrosion is not a separate form of corrosion but shares many
Application to Poultice Corrosion 61
Fig. 2.7 Separation of two glass plates is a model for removing foreign objects from a metal
surface
common features with crevice corrosion, which has been discussed in more detail
elsewhere by the author [4].
The Young and Laplace equation in the form of Eq. (2.8) can be used to
calculate the pressure required to remove a piece of debris from a surface, as shown
in Fig. 2.7. This is because the material properties of the solids are not involved in
the calculation. Knowing this required pressure, we can assess which cleaning
methods will provide sufficient energy to surpass the adhesive force keeping the
debris particle connected to the metal surface.
Accumulated dirt, dust, corrosion products, and other forms of debris often are
contaminated with oil.
Then, the principles of detergency can be applied. See Fig. 2.8 which is adapted
from Adamson [2] and shows an oil-contaminated particle before and after
detachment. The change in free energy for detachment is:
DG ¼ cwo þ cmw cmo ð2:9Þ
where the subscripts refer in turn to the following interfaces: water/oil-contaminated

particle, metal/water, and metal/oil-contaminate particle. For the process to be
spontaneous we need DG < 0. This leads to:
cmo [ cwo þ cmw ð2:10Þ
Fig. 2.8 Surface free energy relationships in the detachment of an oily particle from a metal
surface. Adapted from Adamson [2] with the kind permission of John Wiley
Fig. 2.9 Schematic diagram of a “wet pack” used to test the white rusting (wet storage stain) of
galvanized sheets
Equation (2.10) is the condition which describes the effect of detergents on the
detachment of oily dirt or oily debris from a metal surface. Detergents which can
reduce γwo or γmw will serve to make the inequality in Eq. (2.10) hold. Thus, useful
detergents are those which can adsorb both at the oily particle/water interface and at
the metal/water interface.
Interfacial tensions can be measured with a Wilmhelmy balance in which the
force is measured to immerse a thin plate into a two-liquid system. Details are given
elsewhere for measurements for the chromium/hexadecane/aqueous solution
interfaces [5].
Application to White Rusting
Stacked sheets of galvanized steel can undergo a phenomenon called white rusting
(wet storage stain). Here the spaces between adjacent metal sheets act like capil-
laries for the ingress of water. A differential oxygen cell is set up between metal
under the outside of the flattened water droplet (the cathode) and metal in contact
with the interior of the water droplet (the anode) [4]. White rusting can be mini-
mized by the use of post-galvanized chemical surface treatments, vapor phase
inhibitors, and by storage in a low humidity environment.
The effect of various surface treatments on white rusting can be tested by the use of
“wet packs” of galvanized steel. Here small water droplets are deliberately placed
between stacked sheets for later visual inspection, as shown schematically in Fig. 2.9.
References
1. N.K. Adam, The Physics and Chemistry of Surfaces (Dover Publications, New York, 1968)
2. A.W. Adamson, Physical Chemistry of Surfaces (Wiley, New York, 1990)
3. H. Nguyen, E. Byrd, D. Bentz, C. Lin, Prog. Org. Coat. 27, 181 (1996)
4. E. McCafferty, Introduction to Corrosion Science (Chapter 10). (Springer, New York, 2010)
5. E. McCafferty, J.P. Wightman, J. Colloid Interface Sci. 194, 344 (1997)
Chapter 3
A Recent Model of Passivity
for Fe-Cr and Fe-Cr-Ni Alloys
Passivity is the reduction in chemical or electrochemical activity of a metal surface

due to the formation of a thin protective oxide film tens to hundreds of angstroms
thick.
Fe-Cr Alloys
It has long been known that the introduction of 13 at.% chromium into iron renders
the surface of Fe-Cr binary alloys passive. Various previous explanations and their
limitations have been discussed elsewhere [1].
This chapter summarizes the graph theory model of passivity developed recently
by E. McCafferty [1, 2]. In brief this model shows that there is a critical minimum
concentration of Cr3+ ions in the oxide film so as to form a continuous network of –
Cr–O–Cr– bridges in the oxide film. This condition is provided by the Fe-13 at.%
Cr binary alloy.
A mathematical graph is a collection of points (called vertices) connected by
lines called edges. The mathematical graph for Cr2O3(Go) is shown in Fig. 3.1.
The vertices of the graph represent Cr3+ ions (atoms) whereas edges represent
2−
O ions. The size of Go is given by its number of edges. If the number of vertices
is N, then the number of edges A is A(Go) = (3/2)N (to preserve stoichiometry). The
configuration of Go is given by a quantity called the Randic index X(Go) (3):
X
XðGo Þ ¼ ðijÞ0:5 ð3:1Þ
edges
where i and j are the degrees of the vertices connecting the edge ij. (A vertex has a
degree 2, for example, if two edges meet at that vertex).
It is shown previously that a simple calculation gives X(Go) = N/2.
Next the composition of the oxide film is changed conceptually by insertion of
Fe3+ ions so as to give a new graph G having the new properties A(G) and X(G).

64 3 A Recent Model of Passivity for Fe-Cr and Fe-Cr-Ni Alloys
Fig. 3.1 Mathematical graph for chromium oxide. The vertices of the graph represent chromium
ions and the edges denote oxygen ions. Reproduced by permission of Elsevier
Following Meghirditchian [3] we define a connectivity function
XðGÞAðGÞ XðGo ÞAðGo Þ

d¼ ð3:2Þ
XðGo ÞAðGo Þ
where Go refers to the original graph for Cr2O3 before the introduction of Fe3+ ions,
and G refers to the new graph containing both Cr3+ and Fe3+ ions. Large values of δ
indicate that the connectivity of the new graph is very much different than the
original graph, whereas small values of δ mean that the connectivities of the two
graphs are similar. Note that δ = 0 when G = Go.
Figure 3.2 shows the insertion of Fe3+ ions into the graph for Cr2O3. The
composition of the oxide film after insertion of Fe+ ions is xFe2O3 . (1 − x) Cr2O3.
If D is the number of edges deleted in Go to form the new graph G and one Fe3+ is
inserted per edge deletion (to preserve charge neutrality), then:
Moles Fe3þ D 2x
¼ ¼ ð3:3Þ
Moles Cr3þ N 2ð1 xÞ
Then, A(G) = (3/2)N − D, or
3 5x
AðGÞ ¼ N ð3:4Þ
2ð1 xÞ
Unlike X(Go), the term X(G) cannot be calculated directly because it is not
known which particular edges have been deleted from Go to form G. A stochastic
calculation is employed, the details of which have been given previously [1, 2]. The
result is:
Fe-Cr Alloys 65
Fig. 3.2 Insertion of Fe3+ ions into the graph for Cr2O3. Reproduced by permission of ECS—The
ð1=3Þ5 N(3 5x)

X(G) ¼ 5
f24x4 þ 24ð21=2 Þx3 ð3 5xÞ þ 4ð31=2 þ 3Þx2 ð3 5xÞ2
ð1 xÞ
þ 61=2 x(3 5x)3 þ ð1=2Þð3 5xÞ4 g 1
ð3:5Þ
Insertion of expressions for A(G), X(G), A(Go), and X(Go) into Eq. (3.2) gives
the result (after much algebra):
2ð1=3Þ6 ð3 5xÞ2
d¼ f24x4 þ 24ð21=2 Þx3 ð3 5xÞ þ 4ð31=2 þ 3Þx2 ð3 5xÞ2
ð1 XÞ6
þ 61=2 x(3 5x)3 þ ð1=2Þð3 5xÞ3 g 1
ð3:6Þ
The behavior of δ as a function of the composition of the oxide film is shown in

Fig. 3.3. It can be seen that an atomic fraction of Fe3+ in the oxide film of 0.70
suffices to break down the –Cr–O–Cr– network. Said differently, there is a critical
concentration of 0.30 mol fraction Cr3+ to provide a continuous network of –Cr–O–
Cr– linkages in the oxide film.
From the XPS data of Asami et al. [4] for Fe-Cr alloys in 1 M H2SO4 a cation
fraction of 0.30Cr3+ in the oxide film corresponds to a metal alloy composition of
approximately 13 at.%. See Fig. 3.4. From in situ XANES data of Oblansky et al.
[5] for Fe-Cr binary alloys in an acetate buffer a cation fraction of 0.30Cr3+ in the
oxide film also corresponds to a metal alloy composition of approximately 13 at.%.
See also Fig. 3.4.
Thus, this graph theory approach explains the well-known observation that 13 at.%
Cr induces passivity in Fe-Cr alloys.
The graph theory approach has now successfully explained the occurrence of
critical alloy compositions in 18 different binary alloys [1, 2, 6, 7]. Some examples
are listed in Table 3.1.
Fig. 3.3 The behavior of the

connectivity function δ as a
function of the oxide
composition for
Fe2O3 · Cr2O3 mixed oxides.
Reproduced by permission of
Elsevier
Fig. 3.4 Cr3+ content in the

passive film on Fe-Cr binary
alloys as a function of the
chromium content in the alloy
after immersion (top) in 1 M
H2SO4 [4]. Reproduced by
permission of Elsevier.
(Bottom) After immersion in
an acetate buffer [5].
ECS—The Electrochemical
Society
Fe-Cr-Ni Alloys
This graph theory approach has been extended to Fe-Cr-Ni ternary alloys [8]
because Fe-18Cr-8Ni (304 stainless steel) is a workhorse among the stainless steels.
The approach is similar to that for Fe-Cr binary alloys except that here there are
two foreign cations (Ni2+ and Fe3+) which can disrupt continuous threads of –Cr–
O–Cr– bridges in the Cr2O3 network. See Fig. 3.5. Let xFe2O3 · yNiO · (1 – x − y)
Fe-Cr-Ni Alloys 67
Table 3.1 Some results from the graph theory model of passivity [1, 2, 6, 7]
Alloy Critical mole Calculated Corresponding alloy Observed critical
fraction of cationic mole composition from alloy composition
oxide from δ fraction in oxide surface analysis data from corrosion data
Fe-Cr 0.30Cr2O3 0.30Cr3+ 13 at.% Cr [4, 5] 13 at.% Cr [9]
Ni-Cr 0.17Cr2O3 0.29Cr3+ 7 at.% Cr [10] 8–11 at.% Cr [11]
Al-Cr 0.31Cr2O3 0.31Cr3+ 48 at.% Cr [12] 40–50 at.% Cr [13]
Co-Cr 013Cr2O3 0.23Cr3+ 6 at.% Cr [14] 9–11 at.% Cr [15]
Cr-Mo 0.28MoO2 0.16Mo4+ 21 at.% Mo [16] 20 at.% Mo [16]
W-Zr 0.21ZrO2 0.21Zr4+ 11 at.% Zr [17] 15 at.% Zr [17]
Cr2O3 be the composition of the mixed oxide. For insertion of Ni2+ ions into the
mixed oxide film, (3/2)Ni2+ ions must be inserted on the average for each edge
deletion in order to preserve charge neutrality in the resulting oxide. One Fe3+ must
be inserted for each Cr3+ which is replaced in the continuous network.
Following the procedure used for the Fe-Cr binary alloys the result for Fe-Cr-Ni
ternary alloys is:
Fig. 3.5 Schematic representation of a continuous –Cr–O–Cr– network on an Fe-Cr-Ni ternary

alloy. The oxide also contains Fe3+ and Ni2+ ions. Reproduced by permission of ECS—The
2ð1=9Þ6 ð9 15x 11yÞ2

d¼ fð3=2Þ½2ð3x þ yÞ4
ð1 x yÞ6
þ 3ð21=2 Þ½2ð3x þ yÞ3 ð9 15x 11yÞ
ð3:7Þ
þ ð3 þ ð3Þ1=2 Þ½2ð3x þ yÞ2 ð9 15x 11yÞ2
þ 61=2 ½2ð3x þ yÞð9 15x 11yÞ3
þ ð1=2Þð9 15x 11yÞ4 g 1
For binary Fe-Cr alloys, there is only one variable (x) in Eq. (3.6) and δ has one
root between x = 0 and x = 1. For the case of Fe-Cr-Ni alloys, the root of δ depends
both on x(Fe2O3) and on y(NiO). The solution is a straight line in x and y.
One method of solving Eq. (3.7) is to assume a fixed value of y(NiO) and to
solve graphically for the value of x(Fe2O3) which gives δ = 0. The mole fraction of
Cr2O3 in the oxide film is (1 − x − y).
Figure 3.6 shows the behavior of δ as a function of the composition of the oxide
film for a fixed value of y = 0.20. It can be seen that δ = 0 when x = 0.52 (mole
fraction Fe2O3) and (1 − x − y) = 0.28 (mole fraction Cr2O3). That is a mole
fraction of 0.28Cr2O3 in the oxide film provides a continuous –Cr–O–Cr– network
(when y = 0.2).
This is the minimum amount of Cr2O3 required because larger mole fractions of
Cr2O3 also produce small values of δ. From simple stoichiometry 0.28 mol fraction
of Cr2O3 is equivalent to 0.31 cation fraction of Cr3+, 0.11 cation fraction Ni2+, and
0.58 cation fraction Fe3+. This point lies on the heavy straight line in the triangular
diagram in Fig. 3.7, which shows roots of Eq. (3.6) for additional fixed values of y.
Fig. 3.6 Behavior of the

conductivity parameter δ near
its root as a function of the
Cr2O3 mixed oxide network
which also contains Fe3+ and
Ni2+ ions. (The curve is
shown for a fixed mole
fraction of 0.20 M NiO).
Society
Fe-Cr-Ni Alloys 69
Fig. 3.7 Comparison of

calculated and experimental
compositions of oxide films
on Fe-Cr-Ni alloys [8].
Society
Figure 3.7 also shows experimentally observed compositions of oxide films on

various Fe-Cr-Ni alloys in different environments. The cation fraction of Cr3+ varies
from 0.33 to 0.67, i.e., from the minimum value of Cr3+ required to higher values
indicative of Cr3+ enrichment.
Finally it should be noted that when y = 0, Eq. (3.7) reduces to the case for Fe-Cr
binary alloys, i.e., Eq. (3.6).
References

2. E. McCafferty, Electrochem. Solid State Lett. 3, 28 (2000)
3. J.J. Meghirditchian, J. Am. Chem. Soc. 113, 395 (1991)
4. K. Asami, K. Hashimoto, S. Shimodaira, Corros. Sci. 18(151), 1993 (1978)
5. L.J. Oblansky, M.P. Ryan, H.S. Isaacs, J. Electrochem. Soc. 145, 1922 (1998)
7. E. McCafferty, J. Electrochem. Soc. 149, B333 (2000)
8. E. McCafferty, J. Electrochem. Soc. 154, C571 (2007)
9. P.F. King, H.H. Uhlig, J. Phys. Chem. 63, 2026 (1959)
10. S. Boudin, J.-L. Vignes, G. Lorang, M. da Cunha Bela, G. Blondiaux, S.M. Khailov, J.
P. Jacobs, R.H. Brangersma, Surf. Interface Anal. 22, 462 (1994)
11. H.H. Uhlig, Z. Elektrochem. 62, 700 (1958)
12. E. Akiyama, A. Kawashima, K. Asami, K. Hashimoto, Corros. Sci. 38, 279 (1996)
13. E. Akiyama, H. Habazaki, A. Kawashima, K. Asami, K. Hashimoto, Mater. Sci. Eng. A
A226–A228, 920 (1997)
14. A.S. Lim, A. Atrens, Appl. Phys. A 54, 270 (1992)

15. A.P. Bond, H.H. Uhlig, J. Electrochem. Soc. 107, 88 (1960)
16. P.Y. Park, E. Akiyama, A. Kawashima, K. Asami, K. Hashimoto, Corros. Sci. 37, 1843 (1995)
17. J. Bhattaira, E. Akiyama, H. Habazaki, A. Kawashima, K. Asami, K. Hashimoto, Corros. Sci.
39, 355 (1997)
Chapter 4
Uptake of Chloride Ions and the Pitting
of Aluminum
Introduction
Aluminum and its alloys undergo pitting in chloride solutions. The protective oxide
film present in neutral solutions breaks down in the presence of chloride ions and
pitting ensues.
Aluminum is usually alloyed with various elements to improve its mechanical
strength, and these alloying additions result in the formation of intermetallic
second-phase particles which can disrupt the passive film on the alloy surface.
Chapter 1 considers briefly the pitting of aluminum from the standpoint of an
acid-base phenomenon. More information is given here.
Sequence of Steps
The pitting of pure aluminum occurs in a sequence of several steps [1], as shown
schematically in Fig. 1.17. The first step in the pitting process is the adsorption of
chloride ions onto the oxide-covered surface. It is well known that the outermost
surface of an oxide or an oxide film is covered with a layer of hydroxyl groups (see
Chaps. 1 and 4). For oxide films on aluminum, these hydroxyl groups interact with
protons in solution to produce a positively charged surface:
AlOHsurf þ HþðaqÞ ¼ AlOH2þsurf ð4:1Þ
Then the adsorption of chloride ions occurs mainly through the operation of
attractive coulombic forces between the positively charged oxide surface and the
negatively charged Cl− ion.

72 4 Uptake of Chloride Ions and the Pitting of Aluminum
XPS Studies
Chloride ions next penetrate the oxide film, as shown by X-ray photoelectron
spectroscopy (XPS). Figure 4.1 shows an XPS Cl 2p spectrum for an aluminum
surface held below its pitting potential [2, 3].
Each Cl 2p spectrum is actually a doublet, and there are two distinct sets of
doublets. The lower binding energy doublet was removed by sputtering, indicating
that this doublet was characteristic of surface chloride. The higher binding energy
doublet persisted after sputtering so that this species is due to chloride which has
penetrated the oxide film.
Figure 4.2 shows that the concentration of bulk chloride in the oxide film
increases as the electrode potential approaches the pitting potential. This shows that
the presence of chloride ions in the oxide film promotes film breakdown and pitting.
Previous studies have shown that the pitting potential of aluminum is lower the
higher the chloride concentration in aqueous solution. See Chap. 1. This provides
further evidence that chloride ions promote film breakdown and pitting.
The exact mechanism by which chloride ions penetrate the oxide film is not
known with certainty. There are several possibilities: transport of chloride ions
through the oxide film through oxygen vacancies [4–6], (ii) transport of chloride
ions through the oxide film through water channels [7, 8], or (iii) localized oxide
film thinning or dissolution [2, 9].
Fig. 4.1 XPS 2p3/2-2p1/2 of

chloride on aluminum for a
sample polarized at −0.750 V
versus S.C.E. before and after
sputtering [3]. Reproduced by
Formation of Blisters 73
Fig. 4.2 XPS measurement

of the uptake of Cl− by the
passive film on aluminum as a
function of electrode potential
[2]. Reproduced by
Formation of Blisters
The propagation of corrosion pits on aluminum occurs with the formation and
rupture of blisters at the aluminum/oxide interface [10]. As propagation proceeds
blisters rupture when the hydrogen pressure within the blister becomes sufficiently
large. See Fig. 4.3 which shows a ruptured blister. Figure 4.4 shows a schematic
diagram of the growth and rupture of oxide blisters on aluminum. When the blister
ruptures, the contents of the corrosion pit are opened to the aqueous solution.
Fig. 4.3 A ruptured oxide

blister on aluminum.
Photograph courtesy of P.M.
Natishan and reproduced by
74 4 Uptake of Chloride Ions and the Pitting of Aluminum
Fig. 4.4 Schematic drawing of the growth and rupture of an oxide blister on aluminum
Localized Corrosion Cells
Experimentally determined values for the internal pH and electrode potential E

within several different types of localized corrosion cells are given in Fig. 4.5 [11],
Fig. 4.5 The electrode potential and pH within localized corrosion cells on aluminum
superimposed on a Pourbaix diagram for aluminum. CC refers to crevice corrosion [12, 13],
and SCC to stress-corrosion cracking [14, 15]. The results for an artificial pit on aluminum are due
to Hagyard and Santhipillai [16]
Localized Corrosion Cells 75
where they are superimposed on a conventional Pourbaix diagram for aluminum.

Each of the data sets for the various localized corrosion cells either lie in the region
of corrosion on the Pourbaix diagram or are very close to this region. Data for the
various forms of localized attack lie below the “a” line for hydrogen evolution
showing that hydrogen evolution is thermodynamically possible in each of these
cases of localized corrosion. Hydrogen bubbles have in fact been observed for
aluminum in active crevices [12].
References

2. S. Yu, W.E. O’Grady, D.E. Ramaker, P.M. Natishan, J. Electrochem. Soc. 147, 2952 (2000)
3. P.M. Natishan, W.E. O’Grady, E. McCafferty, D.E. Ramaker, K. Pandya, A. Russell,
J. Electrochem. Soc. 147, 2952 (2000)
4. C.L. McBee, J. Kruger, in Localized Corrosion, ed. by R.W. Staehle, B.F. Brown, J. Kruger,
A. Agrawahl (NACE, Houston, 1974), p. 252
5. C.E. Chin, C.Y. Chao, D.D. Macdonald, J. Electrochem. Soc. 128, 1194 (1981)
6. D.D. Macdonald, J. Electrochem. Soc. 139, 2434 (1991)
7. J.O’M. Bockris, L.V. Minevski, J. Electroanal. Chem. 349, 375 (1993)
8. J.O’M. Bockris, Y. Kang, J. Solid State Electrochem. 1, 17 (1997)
9. T.H. Nguyen, R.T. Foley, J. Electrochem. Soc. 127, 2563 (1980)
10. P.M. Natishan, E. McCafferty, J. Electrochem. Soc. 136, 53 (1989)
11. E. McCafferty, Introduction to Corrosion Science (Springer, New York, 2010), p. 305
12. D.W. Siitari, R.C. Alkire, J. Electrochem. Soc. 129, 481 (1982)
13. N.J.H. Holroyd, G.M. Scamans, R. Hermann, in Corrosion Chemistry within Pits, Crevices,
and Cracks, ed. by A. Turnbull (Her Majesty’s Stationary Office, London, 1987), p. 495
14. J.A. Davis, in Localized Corrosion, ed. by R.W. Staehle, B.F. Brown, J. Kruger, A. Agrawal
(NACE, Houston, 1974), p. 168
15. O.V. Kurov, R.K. Melekhov, Protection Meals 15, 249 (1979)
16. T. Hagyard, J.R. Santhipillai, J. Appl. Chem. 9, 323 (1959)
Chapter 5
Formation of Water Films on the Iron
Oxide Surface
Introduction
One of the earliest quantitative observations in the field of corrosion science was
that there is a critical relative humidity below which atmospheric corrosion is
negligible and above which corrosion occurs [1]. Figure 5.1 shows Vernon’s results
for iron exposed to water vapor containing 0.01 % SO2 [1]. It can be seen that the
corrosion of iron occurs above 60 % relative humidity. The critical relative
humidity is typically 50–70 % for most metals.
This communication reviews previous work by the author on the uptake of water
vapor on α-Fe2O3. This earlier work is discussed here in view of the theme of this
monograph.
Experimental
All experiments were carried out on α-Fe2O3 powder having a nominal B.E.T.
surface area of 10 m2/g. The adsorption of water from the vapor phase was studied
using conventional adsorption isotherms and dielectric relaxation techniques
(known popularly today as AC impedance and usually applied to the solid/liquid
interface rather than the solid/vapor interface). Experimental details are given
elsewhere [2, 3].
The dielectric sample cell consisted of two concentric stainless steel cylinders
between which was packed the α-Fe2O3 powder. Controlled amounts of water vapor
were introduced into the dielectric cell. If the adsorbed water molecule were able to
rotate in the applied alternating signal, then a capacitance and dielectric loss tangent
(tan δ) was measured. The dielectric constant is given by ε′ = C/C0 where C0 is the
capacitance of the empty cell. The dielectric loss ε″ is given by tan δ = ε″/ε′.

78 5 Formation of Water Films on the Iron Oxide Surface
Fig. 5.1 Corrosion of iron in

air containing 0.01 % SO2
after 55 days of exposure
showing the effect of a critical
relative humidity
(approximately 60 %).
Redrawn from Ref. [1] by
permission of the Royal
Society of Chemistry
Results and Discussion
Adsorption isotherms for water vapor at various temperatures are shown in Fig. 5.2.
The number of layers adsorbed is taken from the B.E.T.
Brunauer-Emmett-Teller surface areas and the cross-sectional area of the water
molecule. Coverages of water refer to the number of physically adsorbed water
layers beyond the underlying hydroxyls.
Fig. 5.2 Adsorption

isotherms for water vapor on
α-Fe2O3 showing that
multimolecular layers of
adsorbed water are formed at
a relative humidity of 60 %
and higher [3, 6]. By
Results and Discussion 79
Figure 5.2 shows that at a relative pressure of 0.6 (relative humidity 60 %), the
third layer of water molecules starts to form and at higher relative pressures mul-
tiple water layers ensue. Thus the critical relative humidity for corrosion corre-
sponds to the adsorption of multimolecular layers of water on the oxide-covered
metal surface. If the water layer contains ionic impurities either by leaching from
the oxide film or by incorporation from the atmosphere, then a thin-layer electrolyte
is formed.
Figure 5.3 shows the dielectric constant as a function of coverage for a frequency
of 100 Hz. The first layer of physically adsorbed water does not cause any change
in the dielectric constant. This behavior is due to the existence of a monolayer in
which the physically adsorbed species are held rigidly by double hydrogen bonding
to underlying hydroxyl groups. See Fig. 5.4.
Fig. 5.3 Dielectric constant at 100 c/s as a function of coverage for various temperatures. (Solid
points indicate desorption). Reproduced from Ref. [2] with permission of the Royal Society of
Chemistry
Fig. 5.4 Schematic representation of water adsorption on α-Fe2O3. Reproduced from Ref. [3] with
permission of the Royal Society of Chemistry
The dielectric constant rises sharply with the start of the second water layer. This
increase is due to an increased tendency of water molecules to respond to the
alternating field (Debye dispersion process), so that these multilayers must be more
mobile than the first physically adsorbed layer. This information is also in accord
with the notion that thin layers of adsorbed water start to form at the critical relative
humidity for corrosion.
Typical Cole-Cole plots of ε″ versus ε′ are shown in Fig. 5.5. These plots allow
calculation of the characteristic relaxation frequency fchar [4]. For the formation of
multilayers of adsorbed water on α-Fe2O3, the characteristic frequency is approx-
imately 10–20 kc/s, characteristic of a hydrogen bonded network. Cole-Cole plots
are also called Nyquist plots or complex plane plots.
The tails appearing on the Cole-Cole plots indicate that dielectric processes other
than simple Debye dispersion are operative. They are due to surface d.c. conduc-
tance, and it is interesting to note that they appear when multimolecular layers of
water are formed.
Fig. 5.5 Cole-Cole arc plots

for several coverages of water
vapor adsorbed on α-Fe2O3 at
15 °C. (a) θ = 1.70,
fchar = 176 c/s, (b) θ = 1.82,
fchar = 284 c/s, (c) θ = 2.28,
fchar = 1.09 kc/s. Reproduced
from Ref. [2] with permission
of the Royal Society of
Chemistry
Formation of Liquid-Like Layers 81
Formation of Liquid-Like Layers
The formation of liquid-like films of water on oxide or oxide-covered surfaces

proceeds by the following sequence of events.
1. The first layer of adsorbed water vapor on a bare oxide surface chemisorbs to
form a layer of surface hydroxyls. See Fig. 5.6.
2. The first layer of physisorbed water vapor is immobile and is doubly hydrogen
bonded to the underlying layer of surface hydroxyls.
3. Successive layers of physisorbed water vapor are more mobile.
Supporting evidence for this model comes from immersional calorimetry mea-
surements and from entropy calculations. Details are given elsewhere [5, 6]. The
heat of immersion decreases from approximately 1,000 ergs/cm2 for the bare oxide
surface to 400 ergs/cm2 as the chemisorbed layer of hydroxyls is successively filled
in. As the next layer of water molecules is formed, the heat of immersion decreases
further to 100 ergs/cm2, indicative that this layer is physically absorbed. (This is the
layer of water molecules doubly hydrogen bonded to the underlying layer of
hydroxyls).
The same conclusion is reached on the basis of entropy considerations. The
experimentally determined integral entropy of adsorption is given from multi-
temperature adsorption isotherms to be:

@ ln P DSads
¼ ð5:1Þ
@T / RT
Fig. 5.6 Adsorption of water

vapor on a clean oxide surface
where Φ is the spreading pressure given by:
Z 0
P¼P
/ ¼ RT C d ln P; ð5:2Þ
P¼0
and Γ is the surface concentration of adsorbed water molecules.

The theoretical integral entropy was calculated for two cases: (i) an immobile
physisorbed water layer (i.e. a layer which is doubly hydrogen bonded to the
underlying hydroxyl groups, and (ii) a mobile layer of adsorbed water molecules.
The calculated entropy change upon adsorption includes contributions from
translational, rotational, and vibrational changes. For a mobile adsorbed film, the
adsorbed molecules lose one degree of translational freedom and no degrees of
rotational freedom. For an immobile adsorbed film, the adsorbed gas loses all its
translational freedom and three degrees of rotational freedom if the water molecule
is doubly hydrogen bonded to the layer of underlying hydroxyls. There are con-
tributions due to surface vibrational modes for both immobile and mobile
adsorption. Details of these calculations are given elsewhere [6].
Figure 5.7 compares the experimental entropy of adsorption with the calculated
values for mobile or immobile water layers. It can be seen that the experimental
entropy of adsorption agrees quite well with the calculated values for an immobile
water layer, but not a mobile layer. This lends support to the notion that the first
physisorbed water layer is doubly hydrogen bonded to the underlying hydroxyl
groups.

experimental entropies of
adsorption at 25 °C with
theoretical entropies for
mobile and immobile
adsorbed films of water vapor
[3, 6]. Reproduced by
References 83
References
1. W.H.J. Vernon, Trans. Faraday Soc. 31, 1668 (1935)

2. E. McCafferty, V. Pravdic, A.C. Zettlemoyer, Trans. Faraday Soc. 66, 1720 (1970)
3. E. McCafferty, A.C. Zettlemoyer, Discuss. Faraday Soc. 52, 239 (1971)
4. K.S. Cole, R.H. Cole, J. Chem. Phys 9, 341 (1941)
5. A.C. Zettlemoyer, E. McCafferty, Z. Physik Chem. N. F 64, 41 (1969)
6. E. McCafferty, A.C. Zettlemoyer, J. Colloid Interface Sci. 34, 452 (1970)
Chapter 6
Corrosion Inhibition by Fluorinated
Aliphatic Compounds
Introduction
Using fundamental principles of corrosion inhibition it is possible to “tailor-make”

effective corrosion inhibitors. Much useful information is available from the field of
surface chemistry.
For example, carboxylic acids are surface active agents which adsorb at the
metal surfaces by donation of electrons from the oxygen in the –COOH group to the
metal. When adsorbed from solutions of non-electrolytes, the hydrocarbon tails
are extended outward from the surface to form a close-packed hydrophobic
monolayer. It is well known that surfaces terminating in fluorocarbon groups are
less energetic than hydrocarbon groups and are less wettable than are surfaces
terminating in hydrocarbon groups.
Fluorinated Carboxylic Acids and Amines
If the contact angle θ for a liquid on a solid is small or zero, then the liquid is said to
wet the solid. For large values of θ, wetting does not occur. See Fig. 6.1.
Figure 6.2 compares wettability data for films of fluorinated and conventional
carboxylic acids preadsorbed onto platinum or glass substrates [1]. The wetting
liquid is methylene iodide, an organic liquid often used to rate surfaces. The higher
the contact angle, the less wettable and the lower the energy of the surface.
Figure 6.2 shows that fluorine substitution produces a “Teflon-like” surface with as
little as five carbon atoms.
Thus, the idea of using fluorinated molecules as corrosion inhibitors is intriguing
because of the possibility of in situ formation of a low-energy “Teflon-like” surface.
The actual configuration of the adsorbed inhibitor would not seem important as
either a flat or vertical configuration would present low-energy surface groups to the
solution, as shown in Fig. 6.3.

86 6 Corrosion Inhibition by Fluorinated Aliphatic Compounds
Fig. 6.1 Contact angles for

(top) a wetting liquid and
(bottom) for a non-wetting
liquid on a solid surface
Fig. 6.2 Experimentally

measured contact angles for
water on films of conventional
carboxylic acids and on
fluorinated carboxylic
acids [1]
Fig. 6.3 Fluorinated

carboxylic acids adsorbed in
the vertical and flat
configurations compared with
poly (tetrafluoroethylene)
Experimental Results 87

CH3(CH2)2COOH and
CF3(CF2)2COOH as
corrosion inhibitors for
iron in 3M HCl
Experimental Results
Corrosion rates were determined for high purity-iron wires or high-purity flat
samples using electrochemical polarization curves or colorimetric analysis of the
electrolyte containing freely corroding iron. The corrosive medium was 3M HCl with
and without added butyric acids. Experimental details are given elsewhere [2, 3].
The efficiency % I (% inhibition) of the inhibitor is given in the usual way as:
%I ¼ ðiu iÞ=iu ð6:1Þ
where iu is the corrosion rate of the uninhibited iron (i.e., its open-circuit corrosion
rate) and i is the corrosion rate of the inhibited iron. Figure 6.4 shows that per-
fluorobutyric acid CF3(CF2)2COOH is a poorer—not better—inhibitor than its
conventional analog, CH3(CH2)2COOH.
The reason is that the electronegativity effect of F atoms adjacent to the carboxyl
group is predominant over the potential increase in hydrophobic character. That is,
the strongly electronegative F atoms withdraw electrons from the –COOH group to
make it a stronger acid and weaker base. Thus, the density of electrons available for
bonding with the iron surface is decreased.
This effect is schematically illustrated in Fig. 6.5, which also lists ionization
constants. Thus, the weaker bond strength with the totally fluorinated acid results in
its poorer inhibitive effect.
However, Fig. 6.5 also shows that the presence of methylene groups interposed
between the surface active carboxyl group and the fluorocarbon tail shields the
–COOH group from the inductive effect of the fluorine atom.
These observations lead us to suggest that the most effective fluorinated com-
pounds should be only partially fluorinated, as shown in Fig. 6.6. The chain should
feature several “insulating” –CH2– segments which shield the –COOH group from
the fluorine atoms in the hydrophobic tail.
88 6 Corrosion Inhibition by Fluorinated Aliphatic Compounds
Fig. 6.5 Effect of fluorine substitution on the ionization constant of butyric acid
Fig. 6.6 A “tailor-made” corrosion inhibitor
Table 6.1 shows measured corrosion rates for iron in deaerated 3M HCl with and
without added propyl amine CH3CH2CH2NH2 and pentafluoropropyl amine
CF3CF2CH2NH2. Table 6.1 shows that the partially fluorinated amine is a better
inhibitor than its conventional analog.
It is interesting to note that pentafluoropropyl amine is a better inhibitor than
propyl amine despite the fact that the latter is the stronger base. The beneficial
aspect of an improved hydrophobic tail overcomes the detrimental aspect of a
weaker electron donor.
Table 6.1 Results for the

two propyl amines Compound %I pKa*
CH3CH2CH2NH2 59 10.53
CF3CF2CH2NH2 77 5.7
* Higher values of pKa correspond to stronger bases and better
electron donors
Experimental Results 89
Fig. 6.7 Effect of fluorination and of replacing a –COOH end group with a –NH2 end group on
the corrosion of iron in 3M HCl. Inhibitor concentrations are 1 × 10−3 M in each case
The results for the limited set of experiments discussed here are summarized in
Figs. 6.6 and 6.7. In brief, partial fluorination is better than full fluorination.
Also, amines are better inhibitors than their corresponding carboxylic acids.
Suggestions for Further Work
The effect of systematic progressive fluorination along the hydrocarbon chain

should be investigated. If solubility proves to be a problem, then the fluorocarbon
can be preadsorbed onto the metal surface by adsorption from solution or by
“painting” the fluorocarbon onto the metal surface. Aromatic (cyclic) fluorocarbons
can also be investigated.
References
1. E.G. Shafrin, W.A. Zisman, Upper limits for the contact angles of liquids on solids, NRL report
5985, Naval Research Laboratory (1963)
2. E. McCafferty, Corrosion inhibition by hydrophobing compounds. Paper No. 72, presented at
corrosion/74, Chicago, IL, 4–8 Mar 1974
3. E. McCafferty, M.K. Harvey, Reports of NRL progress (Nov 1976), p. 23

Surface Chemistry

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Surface Chemistry

Uploaded by

Copyright:

Available Formats

SPRINGER BRIEFS IN MATERIALS

ISSN 2192-1091 ISSN 2192-1105 (electronic)

Library of Congress Control Number: 2015932239

Springer Cham Heidelberg New York Dordrecht London

Printed on acid-free paper

Springer International Publishing AG Switzerland is part of Springer Science+Business Media

This communication utilizes a surface chemistry/physical chemistry approach

Alexandria, VA, USA, 2015 E. McCafferty

1 Acid-Base Properties of Surface Oxide Films . . . . . . . . . . . . . . . . . 1

Chelating Compounds as Corrosion Inhibitors . . . . . . . . . . . . .. 33

2 Capillarity and Corrosion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

3 A Recent Model of Passivity for Fe-Cr and Fe-Cr-Ni Alloys. . . . . . 63

4 Uptake of Chloride Ions and the Pitting of Aluminum . . . . . . . . . . 71

5 Formation of Water Films on the Iron Oxide Surface . . . . . . . . . . 77

6 Corrosion Inhibition by Fluorinated Aliphatic Compounds . . . . . . 85

Surface Hydroxyl Groups

© The Author(s) 2015 1

MOHsurf þ Hþ ðaqÞ MOH2 þ surf ð1:1Þ

MOHsurf þ OH MO surf þ H2 O

MOHsurf MO surf þ Hþ ðaqÞ ð1:2Þ

½MO surf ½Hþ

Nature of Acidic and Basic Surface Sites

MOH2 þ surf ! MOHsurf þ Hþ ðaqÞ

3. surface hydroxyls, MOHsurf

MOHsurf ! MO surf þ Hþ ðaqÞ

Fig. 1.2 Various types of

MO surf þ Hþ ðaqÞ ! MOHsurf

6. surface hydroxyls, MOHsurf

MOHsurf þ Hþ ðaqÞ ! MOH2 þ surf

Determination of Isoelectric Points of Metal Oxides

Fig. 1.4 The pH dependence

Streaming Potential Measurement

One method of determining the isoelectric point of an oxide-covered metal is to

rtotal ¼ rsurface þ rdiffuse ð1:6Þ

Contact Angle Measurements

Another method is to measure contact angles of aqueous droplets on the oxide-

Fig. 1.6 Contact angles as

where F is the Faraday, z = 1 here, and Γi refers to surface excesses of H+ or O−.

Wadh ¼ cL ð1 þ cos hÞ ð1:10Þ

and thus are stronger than interactions at the isoelectric point.

Surface Reaction Method of Simmons and Beard

MO surf þ Kþ ðaqÞ MO Kþ surf ð1:13Þ

½MOHsurf þ ½MO surf þ ½MO Kþ surf ¼ No ð1:14Þ

The surface concentration of unreacted negatively charged sites MO surf is

½MOHsurf þ ½MO Kþ surf ¼ No ð1:15Þ

Fig. 1.8 Schematic diagram

MOH2 þ surf þ A ðaqÞ MOH2 þ A surf ð1:17Þ

where h ¼ ½MOHþ 2 A surf =No and is the fractional coverage of reacted

Fig. 1.9 Schematic diagram

Ionic Interaction Model of Parks [26]

MO surf þ 2 Hþ ðaqÞ MOH2 þ surf ð1:19Þ

and the free energy change for Eq. (1.19) is:

DG ¼ 2:303 RT log K ð1:21Þ

Fig. 1.10 Interaction of

Fig. 1.11 A Parks-type plot

XPS Correlation Method

DO ¼ O 1s binding energy 530 ð1:28Þ

Delamar noted that most basic oxides display an O 1s binding energy of

DM ¼ cation binding energy metal binding energy ð1:29Þ

as shown in Fig. 1.12.

Table 1.1 Isoelectric points of some oxides and oxide ﬁlms

Fig. 1.12 XPS photopeak for

Fig. 1.13 Isoelectric points

Ceþ3 þ 3 OH ! CeðOHÞ3 ðKsp ¼ 1:6 1020 Þ ð1:47Þ

i is the electrochemical potential given by: